Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Familiarity, context, and the distinction between literal and nonliteral language Geiger, Odeis 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1993_fall_phd_geiger_odeis.pdf [ 10.25MB ]
Metadata
JSON: 831-1.0098794.json
JSON-LD: 831-1.0098794-ld.json
RDF/XML (Pretty): 831-1.0098794-rdf.xml
RDF/JSON: 831-1.0098794-rdf.json
Turtle: 831-1.0098794-turtle.txt
N-Triples: 831-1.0098794-rdf-ntriples.txt
Original Record: 831-1.0098794-source.json
Full Text
831-1.0098794-fulltext.txt
Citation
831-1.0098794.ris

Full Text

FAMILIARITY, CONTEXT, AND THE DISTINCTION BETWEENLITERAL AND NONLITERAL LANGUAGEbyODEIS GEIGERB.Sc., The University of Oregon, 1965M.A., The University of British Columbia, 1989A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDepartment of PsychologyWe accept this thesis as conformingTHE UNIV SITY OF BRITISH COLUMBIAAugust, 1993© Odeis Geiger, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)eA...--(vThe University of British ColumbiaVancouver, CanadaDepartment ofDate ^/'9 3DE-6 (2/88)ABSTRACTSix experiments investigating the relationship of familiarity andcontext to the processing of metaphor are presented. Experiment1 compares time to understand sets containing idiom or metaphortargets following 1-2 sentence contexts in four conditions:NONLITERAL, where the context was completed by an idiom ormetaphor, PARAPHRASE, where the same context was completed by aliteral target with the same meaning, SURPRISE, where the contextwas completed by a less-anticipated literal target, and LITERAL,where the target from the NONLITERAL condition was used in itsliteral sense in a different context. In Experiment 2, sets ofdead metaphors replaced the idiom sets, and the metaphor sets hadnovel metaphors. Experiment 3 matched the targets in eachcondition for overall printed word frequency, to investigatewhether word familiarity was interacting with type of usage. Italso included an UNFAMILIAR condition, where the same context wascompleted by a much less familiar word used literally.Experiment 4 took 20 contexts from Experiment 3 and askedsubjects to generate their own endings. Experiment 5 replicatedExperiment 3 but with a two-target semantic choice instead of asingle response.^Experiment 6 shortened contexts and reducedtheir information content. Its purpose was to see how muchcontext was contributing to understanding, and whether someconditions would be more affected than others. The results maybe interpreted as indicating that familiarity with the use of aword is important in determining  -speed f^Dea ^ metaphors could be understood just as quickly as words usedliterally, but novel metaphors took longer. Contextualexpectations are also a powerful adjunct to the understandingprocess. When expectations are thwarted, errors andunderstanding time increases. Metaphor understanding isinterpreted as a class-inclusion process in the manner describedby Glucksberg and Keysar (1990), where a word used metaphoricallyis viewed as a prototypical exemplar of a hierarchicallysuperordinate class that becomes extended to incorporate thecontext topic. This process takes time, but metaphors have aresponse latency advantage over surprising or unfamiliar literalwords encountered in context. When context is reduced, metaphorsare still advantageous in terms of time, but are less useful todepth of understanding.TABLE OF CONTENTSABSTRACT^ iiTABLE OF CONTENTS^ ivLIST OF TABLESLIST OF FIGURES^ viLANGUAGE AND MEANING 1THEORIES OF NONLITERAL LANGUAGE COMPREHENSION^ 16INTRODUCTION TO THE EXPERIMENTS^ 32EXPERIMENT 1: PROCESSING IDIOMS AND METAPHORS^ 41EXPERIMENT 2: PROCESSING DEAD METAPHORS AND METAPHORS ^ 61EXPERIMENT 3: THE EFFECTS OF CONVENTIONALIZATION^ 72EXPERIMENT 4: VOLITIONAL METAPHOR GENERATION^ 85EXPERIMENT 5: SEMANTIC DECISION^ 94EXPERIMENT 6: REDUCED CONTEXT 108BETWEEN GROUPS ANALYSIS OF EXPERIMENTS 5 & 6^ 125GENERAL DISCUSSION^ 133CONCLUSION^ 148REFERENCES 149APPENDICES^ 156APPENDIX A - EXPERIMENT 1^ 157APPENDIX B - EXPERIMENT 2 170APPENDIX C - EXPERIMENT 3^ 182APPENDIX D - EXPERIMENT 4 196APPENDIX E - EXPERIMENT 5^ 226APPENDIX F - EXPERIMENT 6 247iv.LIST OF TABLESTable 1^ 52Condition means and standard errors for five experiments.Table 2^ 104Error and time violations for Experiments 5 & 6.V .vi.LIST OF FIGURESFigure 1^ 44Sample idiom item from Experiment 1.Figure 2^ 54Results of Experiment 1 - Idiom and metaphor portions.(N)nonliteral [idiom/metaphor] (P)paraphrase (L)literal(S)surprise.Figure 3^ 65Sample dead metaphor item from Experiment 2.Figure 4^ 69Results of Experiment 2 - Dead and novel metaphor portions.(N)nonliteral [dead/novel metaphor] (P)paraphrase (L)literal(S)surprise.Figure 5^ 77Sample novel metaphor item from Experiment 3.Figure 6^ 79Results of Experiment 3 - Dead and novel metaphor portionsmatched for word frequency. (N)nonliteral [dead/novelmetaphor] (P)paraphrase (L)literal (S)surprise(U)unfamiliar.Figure 7^ 89Sample item from Experiment 4, with responses.Figure 8^ 91Results of Experiment 4 - Comparison of the printed wordfrequency of endings generated by only one subject in a_context_ with the—numbsr—TehbYdted by more than one subject.vii.Figure 9^ 92Results of Experiment 4 - Comparison of mean number ofsubjects making the same response for first generated andlast generated words.Figure 10^ 101Results of Experiment 5 - Dead and novel portions in asemantic decision task. (N)nonliteral [dead/novel metaphor](P)paraphrase (L)literal (S)surprise (U)unfamiliar.Figure 11^ 113Sample novel metaphor item form Experiment 5 and itscounterpart from Experiment 6.Figure 12^ 115Results of Experiment 6 - Dead and novel portions in asemantic decision task with reduced context. (N)nonliteral[dead/novel metaphor] (P)paraphrase (L)literal (S)surprise(U) unfamiliar.Figure 13^ 128Results of Experiments 5 & 6 together - Dead and novelportions in a semantic decision task (N)nonliteral[dead/novel metaphor] (P)paraphrase (L)literal (S)surprise(U)unfamiliar.Figure 14^ 129Error means for Experiments 5 & 6 together - Mean number oferrors per subject for dead and novel portions in a semanticdecision task. (N)nonliteral (P)paraphrase (L)literal(s)qurprice (U)unfamiliar.Metaphor Comprehension -^1LANGUAGE AND MEANINGThe question of how human utterances come to contain andconvey meaning presents psychology with one of its mostintriguing challenges. How is it that we can use language toencode our perceptions, or to recreate others' experiences andbeliefs in our own terms? How is it that every isolated humancultural group invented communication systems based on similarprinciples of named objects and actions, and simultaneouslydeveloped the syntactic structures necessary to make thecombination of these elements meaningful and efficient? How isit that we can recall, store, and understand the meaning of somany individual words so quickly, even if they are used in novelor ambiguous ways? How is it that any normal child can acquireand use such a complex system so effortlessly? In the study ofhuman cognition, psychology has the potential to investigate andilluminate these fascinating issues.The everyday reality of language masks what an extraordinaryprocess it is. It seems simple to say that the words we use havemeanings attached to them; that these meanings are consistentwithin and between individuals in a culture; and that it isthrough this quality of language to permit the sharing ofexperiences in speech and writing that culture and civilizationhave^ ---i-e-t-wherre)irie what meaning is, problemsMetaphor Comprehension -^2immediately arise. Most English speaking people have no troubleunderstanding and using a word like "cute" for example, and yetmany of us would be hard pressed to explain exactly what we aretrying to convey when we use it, or to explicate the commonalitythat would make it an equally appropriate description of ariddle, a teenager, the cut of a dress, a ladybug, a melody, andKevin Costner's behind.While language is the essential attribute that has permittedhumans to pass wisdom along from generation to generation andfrom civilization to civilization, there is some very goodevidence that both thought and memory can exist independent oflanguage. The behaviour of pre-linguistic children plainlyreveals that their abilities to think and remember are already inuse, and their language learning seems to progress from theobservation of an object or an event to the desire to acquire aname by which to identify it. Without at least memory andthought, it would be impossible to make use of words once we areprovided with them, and Fodor (1975) takes this claim one stepfurther and argues that we must have some kind of internal,representational language innately in place in order for us tomake use of the words we learn in our external language ofcommunication. It is indisputably true that we are constantlyextracting meaning from nonverbal symbols and events in ourenvironment. We do not first need a name for something in orderto think about it, and it can be argued that most insight doesnot occur to____us--in—t-lie--foriro a text or an internal lecture. ItMetaphor Comprehension -^3is only when we want to retain, operate upon, or share an ideathat encoding it in written or spoken language becomes useful.The work of Heider [Rosch] (1972) has been ingenious inshowing that language can be distinct from thought. Shediscovered that among the Dani, a primitive culture that had onlytwo color words, people were not constrained to perceive thecolors in their environment in terms of only those two words, butcould make the same distinctions between colors as personsspeaking languages with a much wider range of color terms. Thesefindings decisively countered an earlier hypothesis advanced byWhorf (1956), known as the Sapir-Whorf hypothesis, which statedthat the categories of our language determine the way our worldis perceived and carried the implication that thought andlanguage were synonymous.The Sapir-Whorf hypothesis in its strong form has nowlargely been discredited among psychologists, but a variant of itemerges in a somewhat different guise in the recent work ofGeorge Lakoff (1987; Lakoff & Johnson, 1980). Like Whorf, Lakoffbelieves that the concepts of a language become part of thestructure of thought, and not just its objects. The congruenceof these two theories represents a quintessential example of theslipperiness of language. The Whorf-Sapir hypothesis is oftenidentified as the linguistic relativity hypothesis, since itsupports the idea that reality is viewed relative to one'slinguistic background, and implies that different linguisticgroups might c_ode_reality-d-i-fferentIy:- It is in this way thatMetaphor Comprehension -^4Lakoff incorporates Whorf. Yet the same hypothesis is alsoidentified with linguistic determinism (Tartter, 1986), since thelanguage one speaks can be said to determine the mental view ofreality. This suggests a rigidly categorical view of the meaningof words that seems to be exactly the opposite to what Lakoff istrying to say. If nothing else, such confusions serve as self-reflexive examples in support of a polysemous view of wordmeaning. Just as Lakoff uses Whorf's deterministic theory oflanguage to represent a relativistic model, he makes hisrelativistic model represent a deterministic argument: that ourconcepts arise out of our bodily experience, and that thelanguage we use to express such concepts determines (or at leaststrongly structures) our thought.Lakoff (1987) labels his theory experiential realism. Inhis view, linguistic relativity is important because language isembodied, that is, derived from human bodily and/or social orcultural experience. Since an individual can only know his/herown body from the inside, this makes every individual a uniqueperceiver. Thought and reasoning take the imaginative forms ofpropositions, metaphor, metonymy, and mental imagery. Allconcepts are structured, and new ideas are understood when theyare incorporated into a conceptual structure based on previousexperience. Lakoff calls these conceptual structures IdealizedCognitive Models, or ICMs. ICMs are structured very similarly tothe way Rosch (1973) describes for prototype models, as conceptsorganized • • • • e_ _ -^-cwaities. The ICMs withMetaphor Comprehension -^5membership based on family resemblances parallels the wayconcepts are organized around words. Lakoff bases his beliefthat thought and language make use of the same cognitivestructure on this evidence.One other group of cognitive theories that identifieslanguage with thought are those based on the idea of ParallelDistributed Processing, or PDP (Allman, 1989; Chandler, 1991;Forster, 1990; Hubbell, 1988; McClelland & Kawamoto, 1986). Inthese models, a specific pattern of neuronal activation (or moreaccurately, a simulation of a neuronal activation patterngenerated by a computer) is distributed over a large matrix ofconnection points. Information is contained in the weightings ofthe connections between points, which become strengthened asinformation is processed. Competing units, meanwhile, becomeruled out and inhibited by input from other connections until thenetwork reaches a state of equilibrium. When a connectionistsystem processes language, the strength of the active connectionsare all slightly incremented each time a word is recognized.Word frequency effects are an integral part of PDP models.Similar experiences produce consistent activations of neurons, asis generally described for prototype models, and both learningand generalization of concepts can be explained in this way. Onthe other hand, one limitation of such a model is that since PDPconnections are strengthened by repetition, they present aproblem for creative or nonliteral thinking, since the likelihooding a 'deviant' thought or a new ideaf •enMetaphor Comprehension -^6would be less probable than repeating an established one.Because it explains activation by semantic similarity, such amodel would have a hard time understanding paradoxes orincongruities. In the same vein, it would also appear difficultfor a PDP model to account for lies or sarcasm - cases where aperson is saying one thing while thinking another. A system thatresponds to a pattern of activation presumes that activating aspecific network of neurons could not deal with two simultaneousand contradictory results. While it may be unfair tocharacterize them in such a way, PDP models have some commonfeatures with the 'language is thought' hypotheses of theBehaviorists, which limit what can be known about thought to whatcan be observed through language.Each of the above proposals in some way argues in favor oflinguistic relativity, or at least that the cognitive activityassociated with language processing is inextricably tied tolinguistic input. The other side of this coin is the hypothesisthat the systems of thought and language are separate, and thestrongest proponent of such a hypothesis is Fodor (1975). Fodorargues that thought must have a syntactic structure in the sameway as language does, based on the manipulation of symbols inaccordance with syntactic rules. He claims that you cannot learna language whose terms express semantic properties not expressedin a language you are already able to use.Fodor's (1975) 'Language of Thought' model appears to be theleast real-i-stio- of all -the theories about the relationshipMetaphor Comprehension -^7between language and thought. Language is described by acomputational theory based upon principles of logic and thecombination of representational symbols according to syntacticrules. It demands that words be categorically defined, in orderfor the correspondence between the innate language to remainconsistent with the learned language. Limiting language in thisway precludes either linguistic evolution or linguisticflexibility. Most modern cognitive scientists accept the vastbody of empirical evidence indicating that human behavior isneither logical nor rule governed, and regard such a theory to beunrepresentative of human cognition.Fodor's (1975) system derives from Chomsky's (1965) beliefin linguistic universals, and to be universal, representationsmust be independent of a specific language. Although both allowthat the concepts can be quite primitive, there is still theclaim that we somehow innately possess the elements of all thatwe will ever know. Like Chomsky, Fodor argues that investigationmust be limited to the syntactic, computational properties oflanguage, since the semantic component of language is beyond theability of humans to access or understand.In defense of Fodor and Chomsky, however, is theincontrovertible evidence that all children can learn to uselanguage amazingly quickly and competently, without explicitinstruction. It is also true that very young children can becomefluently bilingual or trilingual, while such an accomplishment isdifficult _and----imper-fect -far - most adults. While it may be goingMetaphor Comprehension -^8too far to argue that we are born with the primitive elements ofall the semantic concepts we will encounter in our interactionwith the world, it does not seem at all unlikely that at leastsome syntactic ability must be innate.Questions of whether or not language and thought areidentical have profound implications for the representation oflanguage. Dictionaries are familiar structures for organizingwords and their meanings within developed cultures, and may beproposed as a likely analogical representation for the waylanguage is stored and accessed in our brains. The problem witha dictionary analog is that it demands a Whorfian conception ofthought - that a word and the meaning associated with it can beregarded as a categorical entity, and that a collection of suchcategorical entities is sufficient to explain how humans uselanguage. The categorical approach to word meaning may be morean artificial tool than a natural capacity. Much currentthinking cites the work of Wittgenstein (1953) and Rosch (1973;1978) with regard to the fact that classical categoriesconsisting of objects and their features do not reflectcategories in the real world, where membership can be graded orfuzzy, and where ad hoc categories that cut across otherwisenonintersecting concepts can be assembled at will (i.e., purplethings; things you can cut with scissors). Dictionarydefinitions represent the epitome of classical categorization andsuffer the same defects when used as a cognitive model of howhuman^ ing.Metaphor Comprehension -^9In my view, the most interesting and plausible model ofderiving meaning from language has its origin in basic levelprototype categories. Such a model would represent concreteexamples derived from experience and/or learned definitions at asubordinate level and abstract themes at a superordinate level.Category prototypes would often be single words, but could alsobe word combinations or phrases. Not all the words that are in adictionary would be represented in prototypes; syntactical wordsare probably stored and manipulated in a completely differentway, and proper nouns may be organized as clusters of distinctivefeatures rather than of similar exemplars. The theorist who Ifeel presents this model best is Schank (1982; 1990), thoughthere have been many types of schema or script models developedthat are all based on similar organizing principles. The majoradvantage of prototype models is that they explain how individualexperiences can be maintained as separate entities, whilesimultaneously accounting for experience-based learning byextracting common elements from recurrent events.The basis of Schank's theory is that knowledge is organizedinto Memory Organization Packets (MOPS), which store events, andThematic Organization Packets (TOPs) which are organized byabstract theme. Schank's MOPs are derived from his earlierscripts, which are short scenarios of the consistent events thatmake up our life experience. The MOPs are built up very much inthe way that Rosch (1973; 1978) describes for prototype models -similar, repeated experiences are stored together and becomeMetaphor Comprehension - 10firmly established, while all of the different features ofindividual episodes, called scenes, remain attached to the partof the script they relate to. The scenes' tendency to becomeattached to different MOPs presents a plausible model for the waywe can flexibly handle ad hoc categories and creative thoughts.Scenes can also function as episodic memories, which we would useto recall a specific instance of the general class of eventrepresented by the MOP. MOPs seem to represent the same level oforganization as basic level categories do, and can be events likevisiting a business office or washing dishes.MOPs are further organized into TOPs or ThematicOrganization Packets. The TOPs are domain-independent themes thatthe prototypical scenes exemplify. One incident may remind us ofanother that is not similar in a concrete way but is similar inan abstract, thematic way. The essential point of this theory isnot literally to describe how memory is encoded but to build onthe principle that every new event that we experience is relatedto the events of our past experience, and that the events areconnected by successive layers of abstraction. Within such amodel we can see the emergence of Rosch's basic level categoriesat the MOP level along with subordinate scenes and superordinateTOPs.There are clear parallels between the conceptual structuresSchank describes and the ICMs of Lakoff (1987). The idea of theorigin of conceptual models in bodily experience is Lakoff's andI think thi (e--itt-t-i-ofound ways of explaining theMetaphor Comprehension - 11innate basis for concepts and thinking. Schank does not addresssuch issues. However, the strength of Schank's theory is that hedraws upon experiences, rather than concepts, as the source ofthought. In Schank's theory experiences become elaborated andcompressed into concepts. In other words, Schank's systemproceeds from experience to thought to language, althoughlanguage is not clearly incorporated within it. Lakoff's systemarises out of experience, too, but the foundational basis of hissystem appears to be organized the other way around, withthoughts emerging from language. Where he initially describesconcepts like anger as emerging from bodily responses likepressure and heat, he has also discussed the grounding ofconcepts in social experience - advancing concepts like"arguments are war" or "theories are buildings". In these uses,the word comes to precede and structure the thought. I view thisas a serious flaw in Lakoff's arguments.Prototype based models in which language and thought areseparate but inextricably bound are finding some support fromneurological evidence - Damasio and Damasio (1992) report thatneural systems in the left hemisphere represent phonemes, phonemecombinations and syntactic rules, and that other left hemisphereneural systems mediate between these structures and those thatcategorize nonlanguage representations - such as sensory andmotor systems - which mediate between the body and itsenvironment. They have found that specific lesions in thesest_ruct_ures can--produ-ce- loss of very specific languageMetaphor Comprehension - 12functions. Damage to the occipital and subcalcarine portions ofthe left and right lingual gyri cause patients to lose both theability to perceive and to imagine color. Lesions in otherregions can produce other highly specific losses, e.g., theability to give proper color names, or the ability to say colorwords. Damasio and Damasio hypothesize that memories aboutexperiences with objects are embodied in ensembles of neurons in'convergence' regions of the brain. When they are activated,many anatomically separate and widely distributed neuronensembles fire simultaneously and reconstruct previous patternsof mental activity. Language can be envisioned as a 'sixthsense' - the input that activates the ensemble. The informationis categorized so that events and concepts can be reactivatedtogether. These conjectures lend support to prototype and schemamodels. For example, their ideas could nicely incorporate thekind of generalization and extension suggested by Shank's (1982)MOPS and TOPs, the internalized cognitive models (ICMs) thatLakoff (Lakoff, 1987; Lakoff & Johnson, 1980) proposes, and thePDP models like Chandler's (1991) as well. According to theDamasios, language performs what P. Churchland (in Damasio andDamasio, 1992) has referred to as "cognitive compression" bycreating categories that reduce conceptual complexity. They thengo on to suggest that because the brain can categorizeperceptions and actions simultaneously along many differentdimensions, symbolic representations such as metaphor can easilyemerge fr_m_thi-s-arehitectu-r-e.Metaphor Comprehension - 13While Schank, Lakoff, and Rosch are all talking aboutconcepts rather than words per se, the line between the two is adifficult one to draw. It seems very likely that when we processlanguage, we are relating what we are trying to understand toconcepts derived from incidents of greater or lesser duration andcomplexity, rather than to verbal definitions, propositions, orfeature lists. Hearing the word 'elephant' may only require thatwe access a scene where we had previously encountered one, whilehearing the word 'beautiful' might require access to a TOPextracted from multiple exposures to events where various thingshave been described or experienced as beautiful. Just as wecompress events into stories, so we might compress stories intothemes, and themes into words. It is worth noting that some ofthe neurological evidence offered by Damasio and Damasio (1992)lends strong support to this possibility. Thought, in such ascheme, is the process of extracting episodes from experience,and language is the way of indexing such episodes. Languagewould then emerge from the same processes as thought, but thoughtwould always underlie, and necessarily precede it. Without somesystem of language, we would have episodes, but no way to extracttheir essence, or to label the result of thematic abstraction.One of the most powerful arguments in favour of thoughtbeing independent of language is the universal existence ofambiguous and nonliteral language. If language determinesthought, then every time we hear a word, the same meaning shouldoccur to_ns--I-f lamrguaq-permits the same word to representMetaphor Comprehension - 14two or more unrelated concepts (bank), related concepts (cold),levels of abstraction (clear), contextually distinguishedinterpretations (sick), and even diametrically opposite meanings(Great news!), then language must necessarily be representingmore than a straightforward correspondence would suggest. Infact, if thought and language were synonymous, we would not needdictionaries at all, since it would be meaningless to talk aboutone word in terms of others. How would we get the words thatseeded the whole procedure? Somewhere we would have to have athought before a word or the system would be one of infiniteregress. A possible solution would be if we accept Lakoff'sexplanation as to the origin of the meaning of words, but hisembodiment idea really only holds for spatial orientation andemotional words that define internal states. It would be hard toimagine how we would metaphorically derive 'lawnmower', 'elf','wrench', or 'hamburger' from either physical or socialexperience without having first encountered an object orexperience, and then, having the concept, acquired a word. Newwords can permit the restructuring of experience and createconcepts where none existed before, but not prior to the elementsforming the new structure being in place. Of course, every newstructure that becomes available then opens itself to includefuture input. A model like Schank's conceptualizes thought andlanguage as independent, and yet does not demand a symbolicsystem of internal representation. Instead the representationtakes the form ot-themes-, gists,_ or stories, based on episodicMetaphor Comprehension - 15events, observed objects, and learned facts that deriveultimately from perceptual processes.The study of nonliteral language usage can uniquely addresssome of the questions associated with the relationship betweenlanguage and thought. Nonliteral language is by definition acontextually dependent use of language, and as such immediatelychallenges theories in which language and thought are identical.It is also incompatible with a theory that requires thought andlanguage to be completely separate systems, since symbolssystematically encode semantic content, and metaphor radicallyalters the content and relationships (especially truthrelationships) of logically structured systems. The study ofnonliteral language calls attention to the pragmatic distinctionSearle (1979) identified between sentence meaning and speakermeaning. It might be said that nonliteral language serves ametametaphorical function in the study of language. Bydemonstrating that the meanings associated with individual wordsare much less structured and specific than they superficiallyappear to be, we come to realize that the same is true of ourapparently literal language as well. With such knowledge, a morerealistic investigation of this remarkable human ability ispossible.Metaphor Comprehension - 16THEORIES OF NONLITERAL LANGUAGE COMPREHENSIONThe earliest philosophers and rhetoricians recognized thatfigurative language had special powers not characteristic ofcommon speech. To Aristotle, writing in the 4th Century, B.C.(in Ortony, 1979) this meant that figurative language was used togenerate an implicit comparison for the purpose of embellishingspeech. In Geoffrey of Vinsauf's treatise on rhetoricallanguage, the Poetria Nova (Nims, 1967), written around 1200, healso identifies metaphor as an instrument of hidden comparison,which is, "introduced not under its own aspect but withdissembled mein, as if there were no comparison there at all, butthe taking on, one might say, of a new form marvelouslyengrafted, where the new element fits as securely into thecontext as if it were born of the theme. . . This type ofcomparison is more artistic; its use is much more distinguished."(Nims, 1967, p. 25). Both Aristotle and Geoffrey regardedmetaphor as a rhetorician's tool that would either embellish ordisguise common language to make it more palatable but lessreliable. Language used in this way was not to be trusted, as itcreated false beliefs by equivocation. Implicit in this view isthat there are always clear, literal words available to expressany meaning, for which metaphor can be substituted if the speakerwants to avoid being direct and (by implication) truthful.___The__same—be4-ief In - Llle prima-dy 61 literal language isMetaphor Comprehension - 17evident in most early explorations into the linguistic nature ofmetaphor, such as Searle's (1979). Searle's intent was todiscover how speaker meaning and sentence meaning "come apart" inmetaphor. He emphasized that there is not one literal meaningand another metaphorical, indirect, or ironic meaning to a singleutterance, but rather only a single meaning, along with thespeaker's intention. The literal sentence sets forth a set oftruth conditions relative to a particular context. He alsodenied that a metaphor is just an inadequate or unbalancedfeature mapping between what Richards (1936) had labelled the"tenor" and the "vehicle" (the underlying idea, and the word usedto convey it, respectively). Instead, Searle derived three stepsfor the comprehension of any phrase, claiming that hearers mustfirst decide whether a special strategy is needed to interpret anutterance, depending on whether it seems defective in its literalform. They then must search for properties in the topic that areamong the values of the metaphorical vehicle, and finally, theymust employ strategies that permit the range of the target to berestricted to that intended by the speaker. These goals ofSearle merge with those of Clark and Lucy (1975) who examined thedichotomy between speaker meaning and sentence meaning in theprocessing of indirect requests. Their three stage model[referred to by Glucksberg (1991) as the Standard Pragmaticmodel] predicts that the primacy of literal language wouldproduce a reaction time disadvantage in the processing ofmetaphors. Clark and Lucy set forth- their model as follows:Metaphor Comprehension - 18First, the listener derives and represents the literalinterpretation of the sentence. Second, he thentests this interpretation against the context to seewhether it is plausible or not. If it seemsappropriate to the context, then it is taken to be theintended meaning, If, however, it does not seemappropriate, either because it contradicts some obviousfact or because it violates a rule of conversation, itis rejected as the intended interpretation. Third, inthe case of such a rejection, the literalinterpretation is combined with an appropriate rule ofconversation, and this leads, by deduction, to theappropriate intended meaning (Clark & Lucy, 1975, p.58).The implications of this model include both that nonliteral usesshould take longer to understand than literal ones and thatderiving a satisfactory literal interpretation would precludeever going beyond it to explore nonliteral implications.Further, the nonliteral interpretation demands a different kindof contextual support. Such clear-cut predictions make the modelparticularly amenable to empirical investigation. The three-stage model depends on the premise maintained by Grice (1975) andSearle (1979) that speech represents a tacit agreement betweenall parties in a conversation to be both cooperative and rationalor else 'appropriateness to the context' could never beunequivocallydeterzdned, Such -a—perspective reinforces theMetaphor Comprehension - 19belief that literal language is always truthful while nonliterallanguage is not.Stage models, while intuitively appealing, have never beenstrongly supported empirically (Pollio, Fabrizi, Sills, & Smith,1984). Hoffman and Kemper (1987), in a thorough review ofstudies that used reaction time to investigate metaphorcomprehension, found no clear evidence in support of such aprocess. Instead, the findings, particularly in the case ofwell-learned nonliteral uses such as idioms and indirectrequests, indicated that these expressions were understood incontext at least as fast as their literal paraphrases (e.g.,Bobrow & Bell, 1973; Gibbs, 1979; 1980).When idioms and indirect requests are found to take nolonger to understand than matched literal expressions, theresults are often attributed to lexicalization. Lexicalizationis the process by which any word comes to be represented in anindividual's vocabulary as a "semantic chunk" (Hoffman & Kemper,1987) with a standard meaning attached. This process appears tooccur with all words we understand and use readily. According toSwinney and Cutler's (1979) Lexical Representation hypothesis,idioms and indirect requests have a stable, communally recognizedmeaning of their own that takes precedence over the compositionalmeanings of the individual words that make them up, accountingfor faster understanding times than would be the case if meaningwere computed by assembling individual word meanings andrelationshigs—Lexicalized-idi-orns-ncome 'greater than the sumMetaphor Comprehension - 20of their parts' in the same way that we ignore the sounds of theindividual letters in a word like "drought" when we becomefamiliar with it. Where idioms are understood as quickly or evenmore quickly than their literal paraphrases, it can be attributedto the 'pre-fab' efficiency of the semantic chunk.The intriguing question raised by the Lexical Representationhypothesis is how the lexicon would handle metaphorical multiplemeanings associated with single lexical entries. It is likelythat very familiar, conventionalized (i.e., dead) metaphors arelexicalized just as idioms are, since their nonliteral meaninghas become stabilized in the same way. To access them, contextmay shape how the lexicon is approached, preselecting among anarray of possible meanings for a single word, or the lexiconmight yield a singular meaning which is shaped by contextafterwards. Swinney (1979) found that immediately afterexposure, both the relevant and the contextually inappropriatemeanings of ambiguous words were facilitated in a lexicaldecision task, suggesting that all known meanings of a word arestored together and accessed simultaneously. Tabossi (1988)reported that prior context could determine whether or not thedominant (most frequent) meaning of an ambiguous word would beactivated. An inadequately constraining context results in theaccessing of both meanings, while a tightly constraining contextprevents access to the inappropriate meaning. Recently, Allen,McNeal, and Kvak (1992) reported results suggesting a wordfrequency advantage-in-lex±cal- decrion tasks, indicating thatMetaphor Comprehension - 21accessibility within the lexicon might be according to a pre-established order based on previous exposure. Their results showthat both phrase familiarity and context can affect how meaningis shaped. Finally, the results of a series of experiments byJastrzembski (1981) using high and low frequency words presentedwithout context indicated that word frequency and number ofmeanings independently influenced recognition speed. Words witha large number of alternate meanings and words with higher wordfrequency counts were recognized faster than words with fewalternates or low frequency. Jastrzembski suggested that eachmeaning of a homograph is represented with a unique logogen.From these diverse studies, it is possible to conclude thateither or both word frequency and word ambiguity informationcould potentially be part of the lexicon, but how the lexicon isordered and how meanings are accessed within it is clearly farfrom resolved. Further, it seems unlikely thatconventionalization can have occurred for novel uses of wordsused metaphorically, where new usages arise spontaneously forwords that have previously been lexicalized with a differentliteral meaning. What might the lexicon contribute to helpingunderstand them?If we accept that lexicalization may contribute to the speedof processing nonliteral language, we are still left with thequestion of why and to what extent the two are related. Ifnonliteral words and phrases can be stored and accessed just likeliteral words, what-accounts for their distinction - i.e., whatMetaphor Comprehension - 22makes a usage unequivocally nonliteral? And if we are usingresponse time as a measure, what can we then say about frequentlyused, conventionalized (i.e., dead) metaphors or infrequentlyused (i.e., unfamiliar) literal words? It seems very likelythat the process of conventionalization is all that separates ametaphor from a literal word. On the other hand, it could bethat literalness and familiarity are simply being confounded,producing results that distort what we might be able to learnabout metaphor. The experiments presented in this dissertationrepresent an attempt to investigate whether theliteral/nonliteral distinction can be separated from familiaritydifferences when response time is used as a measure ofunderstanding.There have been some attempts to explain nonliteral languagecomprehension that do not involve lexicalization. Glucksberg andKeysar (1990) and Glucksberg (1991) discuss metaphor as a type ofclass-inclusion statement. In this interpretation, an "a is b"metaphor produces the attribution of a subset of the features of"b" to "a", so that "a" is consequently included in ahierarchically superordinate category of which "b" is aprototypical exemplar. In this approach, the fact that metaphorsare implicit class-inclusion statements is the cue by which werecognize that a usage is metaphorical. This hypothesis providesa plausible model for how metaphors are comprehended. Inparticular, it suggests how aspects of the literal meaning cancome to be selected_^-14^ e nonliteral meaning of theMetaphor Comprehension - 23same word. Further, it accounts for the familiar observationthat metaphor is often employed when a situation is new, or whenexisting words seem inadequate.Perhaps the strongest advantage of Glucksberg and Keysar's(1990; Glucksberg, 1991) model is that it can be developed toprovide an explanation for the distinction between dead metaphorsand novel ones. Identifying and constraining a new superordinatecategory could take time, but repeated exposure to the same usagewould lead to the superordinate category, named by a prototypicalmember, being available with the same speed as the exemplaritself. The category name would in effect shift to the broadersuperordinate category, encompassing both. For example, if theprimary meaning of 'clear' had been 'visually transparent', useof the word to indicate 'completely understandable' could requirean ad hoc cognitive category construction in its initial use.Following repeated usages, the hierarchical category would becomemore flexible, and eventually both meanings could be subsumedunder the same word. It is not really that there are twodistinct meanings accessed in parallel, but rather a broadersingle meaning that context can shape to its purpose.Another way to explain metaphor comprehension is to placethe responsibility for shaping meaning largely in theexpectations generated by context. This is the approachsuggested by Ortony, Schallert, Reynolds, and Antos (1978) whoconducted two experiments to investigate context effects andImsxsing time differences fbr idioms and metaphors as opposedMetaphor Comprehension - 24to words used literally. In their first experiment, subjectswere given either long and short versions of contexts followed bya target phrase which had a literal interpretation in one contextand a metaphorical interpretation in another. As soon as theyhad read and felt they understood the target, they were to pressa response bar. Ortony et al. found that in these conditions,metaphors and their literal paraphrases showed no statisticallysignificant differences in time to understand target sentenceswhen preceded by long (mean = 45 words) contexts. When contextswere short (mean = 6 words), interpretation time was slower forall conditions, but metaphors were slowed to a significantlygreater extent than literal words. In their second experiment,idioms rather than metaphors followed intermediate lengthcontexts (mean = 24 words) under three conditions. Twoconditions presented the context with a target ending consistingof either the idiom or a literal paraphrase of the idiom, and thethird condition used a different context in which the idiom wasgiven a literal interpretation. The results of this experimentindicated that idioms took no longer to process than either theirparaphrases or the same words used literally. From these data,they concluded that context was the key to understandingmetaphors. As linguistic material is processed, various schemata(Rumelhart and Ortony, 1977) produce expectations that arefulfilled by both literal and metaphorical targets. The maindifference is that there is more consistency in evoked schematabetween context and^ target in the case of literal endings thanMetaphor Comprehension - 25metaphorical ones. The degree of consistency between the evokedschemata of the context and the target accounts for the speed(and, presumably, reliability) of understanding.By emphasizing the importance of context, Ortony et al.(1978) have drawn attention to a factor that must play a vitalrole in constraining the meaning of words used metaphorically.Acknowledging their claim that given inadequate context it isoften difficult to understand metaphors and given substantialcontext they are understood much more readily, the same could besaid about a visually or auditorially degraded word, or acompletely absent one. We can often understand metaphors given avery underdeveloped context ("Some marriages are iceboxes") andat other times the consistent schematas between context andtarget may be very tenuous ("The dawn came up like thunder") andyet the metaphor has considerable cognitive power. Still othermetaphors may generate very consistent expectations and yetconvey little meaning ("My surgeon is a nurse"). It is alsoimportant to realize that often the clarifying context can comeafter a metaphor or from extra-linguistic sources. It does nothave to precede the metaphorical statement to make itinterpretable. Finally and most importantly, the function ofmetaphor in communication is not illuminated by attributing itsinterpretation solely to schemata generated by the context. Whywould we ever choose a word to complete a context that "leavesover" some incompatible aspects, when we apparently have somechoices -AW5ri4-iiferal words that are more fullyMetaphor Comprehension - 26consistent? And why would we ever tack on an obscure statementto reinforce a substantially established one? Any adequateexplanation of metaphor must address these enigmas.Finally, some investigatory metaphor research has beenattempted through the use of connectionist, or ParallelDistributed Processes (PDP) models (e.g., Chandler, 1991;Hubbell, 1988). Chandler (1991) argued that there is nodifference at all in the processing of metaphors and literalwords, though he acknowledges a psychological reality to thedistinction. He cited evidence that metaphors presented incontext show no more processing time than literal controls. LikeOrtony et al. (1978), Chandler emphasized that context must makean important contribution to shaping how any word is understood.What Chandler attempted to convey is that all words are far more'open' than we realize. Even when used literally, ourunderstanding of what is being conveyed is shaped by sententialand extrasentential context to such an extent that we can hear"Get your dog out of the middle of the floor" and let "dog" referto a brown, animate, 30 kg. mammal in one case and a pink,inanimate, 300 g. toy in another. Virtually nothing except avague similarity of shape may unite the two. Metaphorical usesof words, he argues, can be accommodated similarly. Chandlerconjectures that recurring patterns of input features tend tobecome organized into neural patterns of associations, with somecharacteristics having stronger associations due to more constantactivation. Lm-a- simplified summary, metaphors will result inMetaphor Comprehension - 27the strong activation of both the expectations of context and therepresentation of the word used metaphorically in whichincompatible features have been inhibited. No distinctly literalinterpretation is ever made, but literal and metaphorical aredetermined by whether a substantial proportion of microfeaturesin the vehicle need to be inhibited in order to coincide with thedeveloping context.In this dissertation I attempt to build on the strengths ofOrtony et al.'s (1978) Glucksberg and Keysar's (1990) andChandler's (1991) models of metaphor understanding. All threeindependently predict that metaphors as a class can be understoodjust as quickly as literal words as a class. However, making adistinction based on the conventionalization of a metaphoricaluse may affect this prediction. Novel metaphors should takelonger to understand than dead metaphors or literal words,because it should take more time to constrain and construct a newmeaning than to simply apply a category label. The parallel caseto novel metaphors would be unfamiliar literal words, wheremeaning would also need to be constructed from a combination ofwhat might be known and/or available in the word itself, and whatcan be gleaned from context. This effort should likewise berevealed by longer response times.In natural situations, context is a large part of theconstraint process, and a process like the one Glucksberg andKeysar (1990; Glucksberg, 1991) describe could not realisticallyCP/Astra-in, a_metapher untess -uidAt4ief were applied. Although IMetaphor Comprehension - 28will argue that while context probably cannot completelydetermine meaning, deriving the full intended meaning of ametaphor is likewise impossible without considering it. Theability to develop a richer meaning through the combination ofcontextual constraint and familiar meaning provides an efficientand enlightening access to understanding. This dissertationexplores more precisely the contribution of context tounderstanding, both by manipulating the expectations generated bycontext and by manipulating the familiarity of words in the samecontext. Chandler's (1991) connectionist interpretationincorporates contextual effects and could support a model such asGlucksberg and Keysar (1990; Glucksberg, 1991) propose, but it isreasonable to conjecture that a connectionist system shouldrespond more quickly to activations of established connectionsthan to exceptional ones, where a period of interactivecompetition among microfeatures is required for interpretinganomalous input. That is, a PDP system should take more time tounderstand novel metaphors. Both approaches (Glucksberg, 1991;Glucksberg and Keysar, 1990; and Chandler, 1991) acknowledge apsychological reality to the nonliteral/literal distinction thatdraws on the unaddressed features of the metaphorical vehicle.Metaphors are almost universally regarded as adding an ineffabledepth that is not conveyed by a literal paraphrase. This seemsan essential point in regard to understanding why metaphorexists, and how it functions. Even the rhetorical implicationso_f_metaphor - however) appear to lie not in metaphoricity but inMetaphor Comprehension - 29conventionalization. A long-dead metaphor probably has littlemore rhetorical force than its literal paraphrase (which, in manycases, is only a longer-dead metaphor) but a newly introducedmetaphor resonates with the connotations of its source and use.Given the state of understanding about metaphorcomprehension in the psychological literature, a number ofquestions remain unanswered. None of the major models ofmetaphor understanding is without drawbacks. It is my aim inthis dissertation to develop a viable model of metaphorunderstanding that explains and unifies previous results. To dothis I investigate several questions that have shaped the studyof metaphor over the last 15 years, including whether or notmetaphor understanding requires a different cognitive processthan literal word understanding and whether response timedifferences can be considered a valid indicator of such aprocess. A major emphasis is to explore the effects of writtenword frequency (as a measure of conventionalization) on responsetimes for metaphor understanding, and to show how the differencesin 'time to understand' for conventionalized as opposed tounconventionalized words are often, and mistakenly, taken asevidence of a literal/metaphorical distinction. The mostdifficult goal of the dissertation, however, is to begin toprovide an answer to the question of why metaphors exist at all.Specifically, the experiments reported here investigate thefollowing hypotheses:al That-t-he-aeeepLed digtinction between literal andMetaphor Comprehension - 30nonliteral as separate kinds of words is not a viable one. Wordsthat we designate as literal may have origins in other uses thathave become modified over time. Literal meaning is generallyassociated with culturally dominant meaning, and more likelyreflects how frequently users are exposed to a particular usageof the word.(2) That even if the dichotomy commonly recognized betweenliteral and nonliteral is untenable, there is a difference ofdegree between using a word in a conventionalized way andextending the meaning of the word in a way that makes use of someof its prototypical characteristics to expand a conceptual class.The difference is that something new is being conveyed, andreceiving this new information takes time. The compensation isthat through this process, metaphor derives the denotative andconnotative force that contributes to its rhetorical impact.Unfamiliar literal words that undergo the same process will takelonger and suffer in comprehension because the bootstrappingstructure is either considerably weaker or missing.(3) That context is an essential component in constrainingnovel metaphors. Context does not consist solely of the wordsimmediately preceding the metaphor. It may develop from somepreviously established understanding between speakers, somenonlinguistic sensory information present at the time of theexchange, or some words used during the exchange. Withoutcontext, words are likely interpreted according to their mostMetaphor Comprehension - 31(4) That the lexicalized meaning of a word usedmetaphorically is an essential component in deriving meaning fromcontext. Without considering the meaning carried by the literalusage of the metaphor, context has only a stochastic function,allowing us to formulate a prediction of what is to come. Theinformation provided by the following words is alreadyanticipated, and the information conveyed is limited. Usingwords metaphorically goes beyond the expectations generated bycontext to expand and sharpen communication, often in ways thatare thought provoking and enlightening.Although there is some disagreement about the extent towhich context is required for the determination of literalmeaning (see Dascal 1987, 1989; Gibbs, 1984, 1989; Searle, 1979),it seems essential that both the context and the word usedmetaphorically must contribute jointly to understandingnonliteral language. Simply providing a highly developed contextis not enough, nor is the metaphorical word in isolation. Inthese experiments I attempt to demonstrate how the two elementswork together by comparing conditions where contextualexpectations are supported with those where they are violated.Metaphor Comprehension - 32INTRODUCTION TO THE EXPERIMENTSSix experiments are presented. The first compares idiomsand metaphors to their literal paraphrases to pursue the findingsof Ortony et al. (1978) that nonliteral metaphors and idioms donot necessarily take longer to understand than their literalparaphrases. The presentation is divided into blocks ofalternately presented idiom and metaphor portions, and for each,the three conditions used in Ortony et al.'s idiom experiment areadapted for use here. The example of an idiom item provided intheir article (p. 471) is as follows:Idiom versionContext: Dean spoiled the surprise that Joan had beenplanning for their mother's birthday party. Whenhe realized what he'd done, he apologized forhavingTarget: let the cat out of the bag.Literal versionContext: Walking back from the store, Anne found a kittenwhich she put in with her groceries. She got homeand her puppy went wild when sheTarget: let the cat out of the bag.Control versionContext: Dean spoiled the surprise that Joan had beenplanning for their mother's birthday party. Whenhe realized what he'd done, he apologized forhavingTarget: revealed the secret.As the other stimulus items used in their original studieswere not available, an entirely new set of stimuli was developedfor this experiment. Thus, an experiment based on their designw_as_ nPc- sary it order to-establish the efficacy of the newMetaphor Comprehension - 33stimuli. The idiom portion of this experiment was derived fromthe idiom portion of Ortony et al., and the metaphor portion usedthe same presentation technique with single word metaphors. Allcontexts in both portions were similar in length to those ofOrtony et al.'s idiom experiment. Subjects were presented withitems from both portions in alternating sets of 20 items per set.The same context was used for each stimulus item in everycondition except for the LITERAL condition, where the same targetword(s) presented in the NONLITERAL condition were usedliterally. These naturally required different contexts in orderto evoke the literal meanings of the same word(s). The threeconditions adapted from Ortony et al. (1978) and used for bothportions of the experiment were: NONLITERAL, where the target waseither an idiomatic phrase (replaced by a single word deadmetaphor in later studies) or a single word metaphor (replaced bya single word novel metaphor in later studies); PARAPHRASE, wherea literal word target with the same meaning as the NONLITERALtarget was presented in the same context; and LITERAL, where thesame word as the NONLITERAL target was presented in a differentcontext. Sample items are provided with the experiments tofollow, and the Appendices contain a complete listing of thestimuli and targets from all six experiments.In addition to the NONLITERAL, LITERAL, and PARAPHRASEending conditions employed by Ortony et al. (1978), the presentseries of experiments also included a SURPRISE condition in whichliteral word-en-ellilgs that Were plausible and grammatical but thatMetaphor Comprehension - 34violated contextual expectations were included. The SURPRISEendings were primarily used to investigate whether context plusliteral word understanding is inevitably faster than context plusmetaphor, or if the interaction of contextual expectation withtarget word meaning also plays an important part in how readilyunderstanding occurs. It was felt that there should at least bea measurable difference between the PARAPHRASE and SURPRISEresults if the stimuli were valid. As the results will show, thestimuli were extremely sensitive to differences betweenconditions, but the SURPRISE condition proved to be even morevaluable for the additional perspective it provided about therelative effects of context and literal words.The second experiment replicates the first with deadmetaphors and novel metaphors rather than idioms and metaphors.Its importance lies in its ability to demonstrate that it isunrealistic to generalize conclusions about metaphoricalprocessing unless the relative conventionalization of themetaphors is also considered. In order to postulate a specialprocess for understanding metaphors, such a process must work onall metaphors in the same way. If dead metaphors can be shown totake less time to understand than novel ones do, it is untenableto simply conclude that there is a special, time-sensitivecognitive process that interprets metaphors.The third experiment replicated the second one, with twoimportant refinements. First, an UNFAMILIAR condition was addedto both pnrtians- f - which -umtAhiliar literal words with the sameMetaphor Comprehension - 35meaning as the more familiar NONLITERAL and PARAPHRASE targetswere presented in the same contexts. Second, an attempt was madeto equate the overall printed word frequency for all theconditions according to Francis and Kucera's Frequency Analyses of English Usage: Lexicon and Grammar (1982) - except, of course,for the UNFAMILIAR endings, where printed word frequency was keptas low as possible. This experiment was designed to show, first,that unfamiliar words used literally could take at least as longto process as familiar words used nonliterally, and that responsetime results could be explained as a function of theconventionalization of the word with a particular meaningattached, independent of whether the word conveying that meaningwas used literally or nonliterally.The fourth experiment represented a departure from theprevious line. Using the same stimulus items as in Experiment 3,subjects were presented with a sample of 20 of the sentencecontexts and asked to generate their own single word endings.This experiment was motivated by the results of the earlierexperiments, which showed that some metaphors could be understoodas quickly as literal words. It asked whether subjects couldthen also be expected to generate metaphors as freely as literalwords or whether words generated had any relationship to thesubject's familiarity with their meanings. It also looked atintersubject consistency with respect to word generation.Experiment 5 returned to the read-and-respond presentationtonvertry-thb - earlier results in a moreMetaphor Comprehension - 36discriminating response paradigm. Rather than pushing a buttonwhen the meaning of the target was ostensibly 'understood',subjects were presented with two words that had been matched forlength, word frequency and grammatical appropriateness and askedto choose the one that made more sense in terms of the context.In addition to latency, this response paradigm provided dataregarding errors, giving a clearer indication of when targets aremisunderstood. Errors can be as useful as latency informationwith regard to how context and target function together.The final experiment, Experiment 6, reduced context to aminimal level in order to assess how much understanding can beattributed to specific elements of information contained withinthe context, and how much is contained in, or is recoverablefrom, the target word. As it was matched to Experiment 5 withregard to items and presentation, a further series of analysesprovided comparative data for Experiments 5 and 6 taken together.A fundamental theoretical assumption concerning the natureof literal and nonliteral language was made for the purposes ofconstructing the stimuli for this dissertation. Thus far, thewords 'literal' and 'nonliteral' have been used as is typical inresearch on this topic in psychology, which is to say, withoutestablishing any clear definition of them. However, oneintention of these experiments is specifically to investigatewhether or not any differences exist at all between literal andnonliteral and, if so, where the line that separates them shouldbe drawn. In-a---previous - study in which reaction time differencesMetaphor Comprehension - 37based on familiarity were found, Blank (1988) went so far as toidentify four classes of metaphors - very familiar, fairlyfamiliar, transitional, and novel - without specifying anyobjective criterion (beyond their source) for defining theirlimits. In order to impose a working definition by which todistinguish literal from nonliteral for the experiments in thisdissertation, 'literal' is used to denote the most conventionalusage of a word in a cultural sense (i.e., the definitionaccorded primary status in an established dictionary). Allsecondary usages are taken to be nonliteral.The distinction drawn by Searle (1979) between word meaningand speaker's utterance meaning emphasizes the fact that therecan never be a 'stand-alone' metaphor; i.e., one that does notmake use of a word already in the lexicon. The same word cancarry a variety of meanings, but the meanings associated with thewords vary in the frequency with which they are employed. All ofthe meanings associated with a word have some elements in common,but it is possible that only the most conventional usage of theword can be called literal. True homonyms, with completelydifferent etymological origins and independent meanings, such as'bank' as a savings institution and 'bank' as a river's edge,represent a different case. 'Bank' in this case does not haveone literal and one metaphorical meaning. It has two literalmeanings, and each can independently be used as a metaphor.The criterion of conventionality has likewise been given an_ -------definitturF of ---th-e purposes of the experiments in thisMetaphor Comprehension - 38dissertation by equating it with printed word frequency. Thischoice is based on the assumption that literal meanings can beviewed as the ones that are most often conventionalized amongmembers of a linguistic community, although it is recognized thatthere will always be differences in both the number and order oflexicalized meanings among individuals. It is readilyacknowledged that the cognitive lexicon may have both a form anda representation of language that does not reflect frequency ofexposure, or that encodes such information much differently.Word frequency counts were only employed in these experiments asa way to control for the variability that such exposure mightproduce.A cautionary note that underlies all of the research uponwhich this thesis is based is also in order here. The site ofthese experiments, Vancouver, British Columbia, Canada, israpidly becoming a fully multicultural community, and theUniversity from which research subjects are drawn reflects thatcultural diversity. Many students are still in the process ofattaining full fluency in English. Students who learn English asa second language often learn it in a different fashion thannative speakers of a language, and one of the major differencesis that second language learners usually learn syntax andvocabulary independently. Vocabulary is dictionary based, andthus the language is learned in a bottom-up fashion.Conventionalized expressions such as idioms have to be acquiredseparately. Not—surprisingly, these phrases often seem unusualMetaphor Comprehension - 39and can present a great deal of difficulty to new languagelearners. Such evidence strongly favors one of the majorpremises of this dissertation, that idioms and indirect speechacts are learned through exposure and then lexicalized, whilenovel metaphors are not.The experiments presented here represent an attempt to reacha compromise between an awareness of the limitations non-nativelanguage speakers have, and a desire to reflect a representativesampling of the student population. In particular, sincestudents were participating in the experiments for class credit,it seemed unjust to discriminate against some potential subjectson the basis of so elusive a quality as English language fluency.However, it soon became apparent that fluency in English was anessential component to validating the experiments. The solutionwas to note on the signup sheet that fluency in English wasnecessary, and allow students to self-select on that basis.Afterwards, a performance criterion was imposed, both in the data(latency) and post-test (comprehension) portions, to screen outsubjects (regardless of native language) whose scores were twosignificantly beyond the experimental group norm. This took theform of first excluding from analysis any response time under 100msec or over 5,000 msec in Experiments 1, 2, and 3, or over10,000 msec in Experiments 5 and 6. In addition, any individualresponse time more than two standard deviations beyond any .subject's condition mean was not included in that mean, and anysubject whose condition—means exceeded The group condition meansMetaphor Comprehension - 40by two standard deviations in more than half of the conditionswas dropped from the analysis and replaced. These measuresserved to protect the data from persons who claimed Englishfluency but had less familiarity with the language than othersubjects to contribute data. There is no doubt that cleanerresults could have been attained if the experiment had beenlimited to only third generation English language users, butsince such a group no longer represents the social reality ofthis community, the experiments presented here should beunderstood to represent results derived from competent readerswith a wide range of linguistic backgrounds.Metaphor Comprehension - 41EXPERIMENT 1 - PROCESSING IDIOMS AND METAPHORSNumerous empirical studies (e.g. Bowbrow and Bell, 1973;Clark and Lucy, 1975; Gibbs, 1979; Ortony, 1978; Swinney andCutler, 1979) have supported the contention that some forms ofnonliteral language, such as idioms and polite expressions(indirect speech acts), are processed in context at least as fastas their literal paraphrases. Experiment 1 adapted and extendedthe findings of one such experiment (Ortony, et al., 1978) withregard to time for processing different kinds of target endings,and also attempted to isolate more precisely how much subjectsrely on context to shape understanding. The stimuli consisted ofa series of items having .a similar form and presentation to thosein Ortony et al.'s (1978) idiom experiment. In their paper, theresults of two experiments were discussed. The time required tounderstand metaphors was measured in the first experiment wherecontext length was manipulated, while in the second, time tounderstand idioms was measured and context remained constant atan intermediate length. Metaphors, as opposed to idioms, werenot investigated in intermediate length contexts. One aim of thepresent experiment was to investigate whether single wordmetaphors would take no longer than their literal paraphrases tounderstand when context length was controlled, or if Ortony etal.'s finding of no statistically significant difference appliedonly in the case of well ie CiL 1 ed, conventionalized idioms, orMetaphor Comprehension - 42metaphors in well developed contexts. A second goal was toinvestigate Ortony et al.'s claim that context length per se is afactor in how quickly metaphors are understood.The functioning of context was addressed by adding a fourthSURPRISE condition to the three of Ortony et al.'s (1978) idiomexperiment. The SURPRISE condition resembled the situationdescribed by Garrod and Sanford (1988) about the consequences fortext interpretation when role casting assumptions are violated.SURPRISE endings were required to be grammatically correct andplausible, but not as likely to be anticipated from the context.Three reasons motivated the inclusion of the SURPRISEcondition in these experiments. First, its endings were allliteral uses of words displayed in the same context as thePARAPHRASE endings. Thus a direct comparison could be madebetween response times for an expected and an unexpected literalword, to investigate how much of a contribution contextualexpectations could make to understanding when nonliteral usageswere not involved at all. While it is naive to expect that everyword used literally would take the same amount of time toprocess, it is common among researchers when dealing with literaland nonliteral uses of words to treat them as if each examplerepresented an undifferentiated class. By applying thisassumption of literal and nonliteral word homogeneity, processingtime differences between the conditions can be ascribed to thedifferences produced by meeting or thwarting contextual__expectations... Second, reponse time -f6r SURPRISE endings can beMetaphor Comprehension - 43compared to NONLITERAL endings for both metaphors and idioms, tosee the amount of disturbance processing nonliteral words wouldcause relative to unexpected literal words. The third reason wasa purely methodological one, since the predicted result of nodifference between literal and nonliteral targets would onlyconfirm the null hypothesis and leave the experiment open to thecontention that perhaps the stimuli were simply ineffectual. TheSURPRISE condition was expected to slow response times relativeto all other conditions.Each stimulus item consisted of four conditions. These werecreated separately for idioms and metaphors. The conditionswere: NONLITERAL - where the target was either an idiom or ametaphor, PARAPHRASE - where the target was a literal word whosemeaning was similar to that of the nonliteral target word,LITERAL - where the nonliteral word or phrase appeared in adifferent context that required it to be understood literally,and SURPRISE - where the target was a literal word that was notlikely to be anticipated (i.e., was surprising) given thecontext. Targets within a stimulus item were matched as closelyas possible for length. Contexts for each stimulus in all butthe LITERAL condition were identical. The dependent variablethroughout the experiment was the response time to understand thetarget ending, given unlimited self-paced exposure to theprevious context. An example of an idiom item from Experiment 1is shown in Figure 1.iT^the-founcandItions Within each portion allowed aMetaphor Comprehension - 44CONTEXTChris and his friends were fooling around when theyaccidentally broke his mother's treasured crystal vase.Several days passed before his mother noticed, and by thenChris was relieved toENDINGS [NONLITERAL] get it off his chest.[PARAPHRASE] confess his guilt.[SURPRISE] get his cuts treated.LITERAL CONTEXTElmer was working on his car when the jack slipped,pinning him underneath. He was not seriously injured, butit took three rescue workers toENDING[LITERAL] get it off his chest.p e iom item from Experiment 1.Metaphor Comprehension - 45number of useful comparisons of response times. ComparingNONLITERAL to PARAPHRASE provided insight into how long it takesto understand the same meaning when expressed nonliterally andliterally. Comparing NONLITERAL to LITERAL gave a measure of howlong it takes to understand the same word or phrase usednonliterally and literally. Comparing PARAPHRASE to SURPRISEprovides a contrast of how long it takes to understand a literalword or phrase with a likely meaning and a literal word or phrasewith an unlikely meaning in the same context. And the NONLITERALto SURPRISE comparison examines the relative disruption createdby two possibly different cognitive effects.MethodSubjects Thirty-two undergraduates were recruited from the Universityof British Columbia Psychology Department subject pool. Eachreceived one class credit for an hour of participation. Allindicated that they were competent readers and speakers ofEnglish. Two subjects were dropped and subsequently replacedwhen a preliminary tabulation indicated that their overallresponse rates were more than two standard deviations slower thanthe combined means for all subjects in over half of theconditions.Stimuli and ProcedureSubjects were given an introduction sheet to explain theexperiment before it began. They were seated in front of a_Ifewl tf___Packard Neetra- ES-7-12 - computer with a standard colorMetaphor Comprehension - 46monitor in monochrome mode and introduced to the responseprocedure. A mechanical push button connected to the computertimed the responses to within one millisecond (msec) accuracy.Five unrecorded practice trials preceded the data collection.Subjects were told that a post-test consisting of a written,true/false recall questionnaire would follow the visualpresentation so it was important to actually read and understandthe items, not just to respond quickly. To ensure thatcontextual effects were consistent for each presentation, it wasstressed to subjects that the success of the experiment dependedon their reading a context each time it recurred, even if it wasrecognized from a previous trial.Idiom and metaphor items were constructed in the same way,with similar length contexts of one or two sentences for bothtypes of item. Contexts varied in length from 22 to 39 words inthe case of idioms and 18 to 27 words for metaphors. The idiomtargets were 3 to 7 words in length, while metaphors weregenerally a single word. All stimulus contexts and targets, anda copy of the post-test, are provided in Appendix A. Subjectswere presented with an entire context at once, which remained onthe screen until they had read and understood it. They thenpressed a button to display the target. When they had read thetarget and understood its meaning in terms of the sentencecontext that preceded it, they pressed the same button as quicklyas possible. Twenty different idiom items and twenty differentmetaphor it_ens_were-presented7- ach-hAVIE4 four differentMetaphor Comprehension - 47targets, for a total of 160 trials per subject. Trials weregrouped into eight sets of twenty stimuli with either idiom ormetaphor items as targets (4 sets of each). The order of setswas counterbalanced, with each subject receiving a differentorder, but idiom and metaphor sets always alternated. Each sethad one condition (e.g., NONLITERAL, PARAPHRASE, LITERAL,SURPRISE) of each item, and an equal number of the same condition(5 of each). The order of presentation of items was randomizedwithin each set. After all the responses were made, subjectswere given a post-test consisting of a paper-and-pencilquestionnaire with 32 true/false items to be recalled from the160 presentations. The items were designed to test whether aspecific meaning of the completed presentation was retainedcorrectly.Corrections for violations of the statistical sphericityassumption for repeated measures analyses of variance (ANOVA)were calculated for all data reported here and in the experimentsthat follow. Whenever degrees of freedom or significance levelsare reported in any of the following experiments, they reflectvalues corrected by applying a Huynh-Feldt epsilon factor to theexperimentally derived degrees of freedom (Kirk, 1982).The SURPRISE condition represented a somewhat differentsituation than the three other conditions, which shared eithercontext or target with one another and so directly addressed theliteral/nonliteral question. Its longer response times wereenotglh to pr^o- • - icily violations, and it was feltMetaphor Comprehension - 48that a more meaningful interaction analysis would involve justthe NONLITERAL, LITERAL, and PARAPHRASE conditions. These arereferred to as the base conditions. Two primary analyses - onefor all conditions and one for only the three base conditions -are reported in all the following experiments, but subsequentanalyses are for base conditions only. Throughout, responselatencies for all the conditions are supplied in the tables andillustrated in the figures.The conditions in these experiments constitute a within-subjects factor, so that the same context is repeated three timesfor each subject (four times in Experiments 3, 5, and 6) with adifferent target each time. The decision to repeat contexts wasmade because it was felt that inter-subject variability in theunderstanding of nonliteral language was likely to be very highin experiments of this type, and a repeated measures analysiswith a sphericity correction would be a better way to controlerror variance than a between-subjects design. The tradeoff inthis choice is that with repeated presentations of context,priming effects can occur. Since order of presentation ofcontext sets was different for every subject, it was felt thatany priming effects would be distributed evenly over conditions.However, as an additional check against the possibility of sucheffects, analyses by order of presentation were performed. Thiswas done by combining all the first, second, third and fourth(where applicable) exposures to a context type, and calculating amean score for each -then applying the same criterion ofMetaphor Comprehension - 49deviation as for condition means, by eliminating all scores under100 msec, over 5,000 msec (10,000 msec in Experiments 5 and 6),or beyond two standard deviations. ANOVA was then calculated todetermine whether there were any significant differences betweenfirst exposures and last exposures to the same context in anycondition. Order ANOVAs are reported only for the three baseconditions.Although the consistent finding of no significantdifferences in order effects in this and subsequent analyses canbe challenged as indicative of lack of power in the test, visualinspection of order means lends support to the premise thatpriming as the result of order of presentation have not createdspurious effects in the data. First presentations appear to showslightly longer response times in most of the conditionsanalyzed, but this effect is generally consistent over thevarious conditions and the pattern of the data is consistent withthat of the overall means for both the first and lastpresentations. In any event, presentation order could not haveproduced condition differences, since order of presentation wasrandomized over conditions and between experiments. Further, theconsistently faster responses shown by the LITERAL condition(where contexts appeared only once) when compared to thePARAPHRASE condition in all but Experiment 6 make it unlikelythat priming can be cited as an explanation for the effectsobtained in these experiments. Results throughout report onlyorder_by portion  by -cond±tturr interactions.Metaphor Comprehension - 50Finally, the data were also analyzed by dividing thestimulus contexts in each experiment according to theirarbitrarily assigned numbers into groups of even and odd, andperforming an analysis of variance with these groups as a factorto provide some assessment of stimulus generality. Outliers weredropped from these groupings in the same manner as in the orderanalysis. While it must be stressed that the nature of stimuliof this type is inherently singular, it is true that the stimulishould be at least generally representative, and this seemed areasonable check to impose. ANOVAs by half (even/odd) arereported only for the half by portion by condition interaction,and only for the three base conditions.Results and DiscussionResults on the post-test for all subjects ranged from 81-100% correct, with a mean of 91.5%, indicating that subjects hadread the contexts of all conditions to the point ofunderstanding.Responses for any individual target that differed by morethan two standard deviations from that individual's mean for thecondition were dropped and the mean recalculated based on theremaining responses. Responses of less than 100 msec or morethan 5,000 msec were also eliminated. These discarded responsesincluded obvious keypress errors at the low end, and cases wherea target was either incomprehensible or unknown to the subject(as in the case of idioms) at the high end, and amounted to lessthan 5% of the _total data. -Neap -response times in msec for everyMetaphor Comprehension - 51condition of this and every subsequent experiment are provided inTable 1, along with their standard errors.A 2 X 4 repeated measures analysis of variance was conductedon the mean response times for all items in the four conditionsfor both portions (idioms and metaphors). The two-portion, four-condition overall ANOVA revealed a statistically significanteffect for portion [F(1,31) = 171.45, p < .001] and condition[F(1.4,42.6) = 46.55, p < .001], and a significant portion bycondition interaction [F(2.6,80.3) = 17.47, p < .001]. Thestatistical significance of the portion result is attributable tothe reading time differences, given the fact that the metaphorswere mostly single words, while the idioms were several wordslong. Mean response times obtained in each condition appear inFigure 2.In the three-condition analysis, the effects of portion andcondition are significant at the same level as for the four-condition analysis, but the interaction of portion and conditionis no longer statistically significant when the SURPRISEcondition is excluded [F(1.9,58.4) = 1.35, p = .267]. Theseresults do not support the conclusions of Ortony et al. (1978)and others, that idioms and metaphors presented in context takeno longer to understand than their literal paraphrases or thanthe same words used nonliterally. As Figure 2 shows, responsesin the NONLITERAL condition for both portions are consistentlyslower than for the PARAPHRASE and LITERAL conditions.Possible_pr-ia-i-ng-ef-fects—dUd- tO repeated presentations ofTABLE 1.CONDITION MEANSMetaphor Comprehension(IN MSEC) FOR FIVE EXPERIMENTS-^52PORTION CONDITION EXP 1 EXP 2 EXP 3 EXP 5 EXP 6DEAD NONLITERAL 1092* 765 844 1115** 1144**DEAD PARAPHRASE 1029* 780 849 1197** 1178**DEAD LITERAL 1041* 730 766 1024** 1138**DEAD SURPRISE 1341* 998 963 1675** 1475**DEAD UNFAMILIAR 1165 1592** 1648**NOVEL NONLITERAL 783 936 922 1503** 1589**NOVEL PARAPHRASE 719 805 828 1136** 1182**NOVEL LITERAL 688 737 818 1049** 1349**NOVEL SURPRISE 840 940 1017 1848** 1546**NOVEL UNFAMILIAR 1214 1767** 1789**STANDARD ERRORS (IN MSEC) FOR FIVE EXPERIMENTSDEAD NONLITERAL 51.8*^38.0 59.9 39.1** 31.5**DEAD PARAPHRASE 43.5* 35.4 56.8 55.7** 42.0**DEAD LITERAL 47.4* 34.1 52.5 39.8** 51.9**DEAD SURPRISE 72.7* 61.5 78.1 99.0** 72.9**DEAD UNFAMILIAR 94.2 94.6** 70.4**NOVEL NONLITERAL 38.4 59.0 71.5 112.3** 71.1**NOVEL PARAPHRASE 31.3 37.7 52.2 38.5** 36.9**NOVEL LITERAL 29.7 35.2 58.8 35.3** 55.9**NOVEL SURPRISE 44.5 56.9 80.1 147.4** 79.2**NOVEL UNFAMILIAR 105.9 117.4** 74.7*** Idioms rather than dead metaphors.** Semantic _decis-i-on- rather- than time to understand.Metaphor Comprehension - 53the same stimuli were investigated in the order analysis. Thisanalysis looked at whether response times for first-presentedcontexts differed significantly from times for last presentedcontexts. The ANOVA of the three base conditions found nostatistically significant order by portion by conditioninteraction [F(1.7,51.4) = .94, p = .381].Generalizability of contexts was investigated by comparingresponse times for even numbered contexts to those for odd-numbered ones. In this analysis, a significant group by portionby condition interaction was found [F(1.4,44.3) = 5.16, p =.018]. Odd numbered items were consistently faster than evennumbered items for both portions, even though the portions werecompletely independent. While no explanation can be made toaccount for this result, it does sound a warning about thevagaries of linguistic stimuli, and the need to obtain strong,consistent effects before any conclusions can be drawn.Generalizing over the results of both of their experiments,Ortony et al. (1978) concluded that "What determines thedifficulty of processing is not literalness but relatedness tocontext" (p. 475). The present results qualify such acontention. While literalness does not produce a unilateraleffect, since response times to the same context were bothshorter (in the PARAPHRASE condition) and longer (in the SURPRISEcondition) than NONLITERAL responses, it does appear to have areliable effect. The base condition analysis shows astatistically significant effect for condition [F(1.9,58.4) =Metaphor Comprehension - 54 MEANS - EXPERIMENT 11400Z 3(.1) 12002ILI 1000z0tx 800600METAPHORI IDIOMCONDITIONS - EXPERIMENT 1Figure 2: Results of Experiment 1 - Idiom and metaphor portions.(N) nonliteral fidio_mimetaphori (P) paraphrase (U literal (S) surprise.Metaphor Comprehension - 5524.32, p <.001].While responses take longer when the context/targetrelatedness is disrupted in the SURPRISE conditions, just asOrtony et al. contended, slower response times also occur for theNONLITERAL conditions. An examination of the data of Ortony etal.'s (1978) metaphor experiment suggests that their data show asimilar pattern to those obtained in the present experiment,despite marginal statistical results indicating no significantdifference in their long context condition [F(1,66) = 2.73, p >.10].^A rough estimation derived from their Figure 1 graph (p.470) indicates that even when metaphorical phrases were presentedfollowing a long context they took approximately 250 msec longerthan their comparable paraphrases. Their long context conditionproduced an estimated ratio of difference/overall mean responsetime results of 250/1950 msec, or 12.8%, while the samecomparison for the metaphor portion of the present experimentproduced a ratio of 95/735 msec, or 12.9%. It is possible thatthe only reason the differences were not statistically reliablein Ortony et al.'s experiment was that reading an entiremetaphorical phrase produced more variability than reading asingle word metaphor. In addition, variability in their resultswas higher because analyses were based on between-group sampleswith only 16 trials per group.The findings of the idiom portion of this experiment arealso at odds with Ortony et al.'s (1978) conclusions. They foundno _significant-difference- between idioms and controls (NONLITERALMetaphor Comprehension - 56and PARAPHRASE conditions here), while LITERAL uses of idiomstook longer. In the idiom portion of the present experiment, theNONLITERAL condition took somewhat longer than the PARAPHRASE orLITERAL conditions, although when an ANOVA followed by Tukeytests was performed on the idiom portion alone, these differenceswere not significant. It is possible that the idioms selectedfor this experiment included more that were unfamiliar tosubjects than had been the case in Ortony et al's experiments, oralternatively, that the subject population was less knowledgeableas to their use. The overall conclusion from comparing resultsacross experiments is that the high variability produced by usingidioms and metaphorical phrases does not effectively elucidatewhat is occurring in nonliteral language processing.In outlining the hypotheses underlying Ortony's et al.'s(1978) experiments, predictions were made concerning the effectsof contextual support on understanding time for metaphoricalphrases as opposed to literal ones. An inverse relationship wasproposed, with longer contexts producing shorter response times.These claims are understood to be offered ceteris paribus.However, this is not always the case, since familiarity withspecific usages must also play a significant part in how readilyany meaning is understood. Idioms are likely to have aconsistent, easily accessible meaning, while metaphoricalmeanings can vary from novel to commonplace. Since it is possiblethat familiarity exerts a differential effect over response timeat least as _p_olgertu-l-a-s---tirat---af-the --difference between matchedMetaphor Comprehension - 57metaphorical and literal phrases by Ortony et al.'s definition,it is useful to emphasize the contribution of word familiarity.Accepting their argument concerning the relationship betweencontext and comprehension at face value, the SURPRISE conditionin the present experiment should take no longer than thePARAPHRASE condition, because the context is identical in bothconditions, and the word or phrase that completes the SURPRISEcondition is always used literally. Ortony et al. predict thatmetaphors can require more time for comprehension than literalphrases when there is little contextual support and can be justas fast as literal phrases when context is adequate, but such aprediction can only be made against an assumption that theclasses of literal words and metaphors are not affected by anyother source of variance. Clearly this is not the case for thepresent experiment. Despite its having a literal word target,the SURPRISE condition takes significantly longer than all otherconditions for both the idiom and metaphor portions. Suchevidence strongly suggests that we must consider the effects ofthe target as well as the context if we are to gain a realisticinsight into the processing of nonliteral language.Context has a powerful shaping effect on expectations, andthese expectations are thwarted in the case of SURPRISE targets.This result is consistent with the approach of Garrod and Sanford(1988), who suggested that text interpretation occurs against aconstantly changing model of meaning that includes backgroundknowledge. In thelr model, die entities mentioned in the textMetaphor Comprehension - 58are represented in an "explicit focus" partition of memory thatbecomes mapped in an "implicit focus" partition by role pointersthat encode customary expectations associated with them in"scenarios". They suggest that readers will encounter difficultyprocessing sentences that violate these implicit assumptions.The SURPRISE condition results support this claim.Language is not always predictable, and no matter how welldeveloped a context is, there is always a range of possible waysa sentence or idea can turn out. A long context may producefaster response times because it constrains the possibilitiesmore, but there is no necessary relationship between degree ofconstraint and metaphor processing. On the basis of responsetime alone, it appears that manipulating relative expectancy forliteral words can produce response time differences that may beeven greater than the differences that metaphors produce, anddoes not isolate what is unique to metaphor.Kintsch and Mross (1985) have reported that thematicallyappropriate words embedded in a text do not prime theidentification of other words (in terms of both sense activationand sense selection) unless they are also associatively related.They conclude from these data that word identification is likelyto be controlled by encapsulated input modules such as thosedescribed by Fodor (1983), that would make them impervious totop-down thematic context effects. Such a claim would mean thatsingle word metaphors would be impossible to understand, unlessthe intended metaphorical meaning were somehow encoded within theMetaphor Comprehension - 59module prior to use. If the PARAPHRASE and SURPRISE conditionshad shown no difference, it is possible that an explanation likeKintsch and Mross's might account for the data. However, theirresults actually showed statistically significant thematicpriming in 3 of their 4 experiments. Further, they claim thatsuch effects only occur if test words follow priming wordsimmediately, "before the thematic context has had a chance tosettle upon a particular sense of the homograph priming words".Their claim then, is not that thematic context has no effect, butonly that it does not determine sense activation. The presentexperiments are designed to investigate the full process ofunderstanding. Even Fodor would allow that for understanding tooccur, thematic context must be involved at some stage of theunderstanding process.Metaphors, including those in the present study, differwidely in familiarity of use. Most dead metaphors become soconventionalized that they appear as secondary definitions indictionaries. Other metaphors, whether they arise spontaneouslyin conversation or are produced as the result of consciousdeliberation, are recognized as highly novel. Unless they becomepopularized and move into the cultural mainstream, they representunique usages. In either case, a word is being used to conveysome meaning that is not its recognized literal meaning. It isimportant to realize that even though dead metaphors are moreconventionalized, they are not any less metaphorical in adefinitional_- sense -t-h-a-rr-TroVre-i—me-t-aphors. Most research onMetaphor Comprehension - 60metaphor makes no distinction on this criterion, and metaphorsused experimentally appear to be more often of the dead, orhighly conventionalized type. The present experiment itselfconfounds this distinction. Its idioms were considered to befamiliar to most subjects, but its metaphors were not evaluatedwith respect to familiarity, and were far more often dead thannovel. A statistically more powerful examination could beperformed if single word dead metaphors were studied rather thanidioms, and if these results could be contrasted with those ofhypothetically unlexicalized novel metaphors. This observationmotivated the second experiment in this series.Metaphor Comprehension - 61EXPERIMENT 2 - PROCESSING DEAD METAPHORS AND NOVEL METAPHORSDead metaphors are conventionalized metaphorical usages.According to Kittay (1987) dead metaphors are ones for which themetaphorically transferred term remains the only term in thelanguage which expresses that concept. She explains that deadmetaphors often result from catachresis, where a word isdeliberately adapted to fit a previously unnamed concept and thenbecomes the sole referent, as in the use of the word 'current'with reference to the flow of electricity. Kittay also allowsthat dead metaphors may derive from unique metaphorical usagesthat have become widespread and common over time. She regardsmetaphors transformed in this way as standard uses, but stopsshort of equating them with literal words. This is the criticalquestion this experiment seeks to investigate.Conventionalized metaphors have often been acknowledged tobe a special problem for the study of nonliteral languageprocessing. Dascal (1987), perhaps the strongest recentproponent for the existence of literal language, has allowed thatindirect speech acts, idioms, and 'frozen' metaphors can beconsidered literal as much by virtue of their conventionality astheir compositionality. He regards ironic utterances and novelmetaphors to provide a more convincing test of whether literalmeaning plays a part in nonliteral understanding. In particular,he cites ^•:, experimen s as failing to recognize thatMetaphor Comprehension - 62highly conventionalized usages may produce results that areindistinguishable from those of literal expressions. At the sametime, however, Dascal still claims that literal meaning is alwaysaccessed regardless of context. Searle (1979), while allowingboth that rejection of literal meaning takes place only if anutterance is "radically defective" and that dead metaphorseventually become literal, still contends that "Sam is a pig",when used with a human referent, would be an example of aradically defective utterance if taken literally. Accessing andrejecting the literal meaning of 'pig' would appear to be anecessary step before a metaphorical understanding is constructedin his model. Overall, there seems to be considerable confusionas to how conventionality would affect predictions concerningmetaphor.Since dead metaphors represent the sole or most familiarword for a concept, but not the sole or most familiar usage of aword, they resemble both literal words and metaphors. However,in their role as metaphors, they should evoke the same cognitiveprocessing demands as any other metaphor, unless theirlexicalization is sufficient to transform them into literalwords. Such a finding would certainly be problematic for three-stage models like Clark and Lucy's (1975) or Searle's (1979), andwould require that any research on metaphor understanding atleast direct some attention toward establishing criteria forlexicalization.Experiment 2 was designed to investigate whether novelMetaphor Comprehension - 63metaphors would take longer to understand than dead metaphors,and, if so, whether this result could be taken as an indicationthat the novel metaphorical meanings had not yet becomelexicalized. Three alternative results were possible. Ifneither dead nor novel metaphors took longer than literal words,it could be assumed that there was simply no difference betweenunderstanding literal and nonliteral words when preceded byadequate context. Finding this result would lend support to theconclusions of Ortony, et al. (1978). If both took longer, oneexplanation could be that all nonliteral understanding entails adistinct process of metaphorical interpretation, as the three-stage models such as Clark and Lucy's (1975) seem to imply. Andif dead metaphors took no longer to understand than literal wordsbut novel metaphors did take longer, lexicalization of a specificusage as a result of familiarity, rather than any specialattribute of metaphor understanding, would appear to beoperative. Blank (1988) reported evidence consistent with such ahypothesis in his study. As in the previous experiment,intermediate length contexts of 1 - 2 sentences were used, andthe same contextual effects as were examined in Experiment 1 werealso assessed here, for the same reasons.It is essential to keep in mind that the literal targets inthese experiments were words that were judged unequivocally to beconventionalized, and that what we define as novel metaphors arebased on singular uses of exactly the same words. I have definedconventionalization to be a function of a consistent meaningMetaphor Comprehension - 64being attached to a word and not of the word itself, so that ahighly conventionalized meaning and an unconventionalized meaningcan both be represented by the same word, with an array ofpossibilities in between, that only context can clarify. Theformer is what has been designated a literal usage and the latteris called a novel metaphorical usage. The LITERAL and NONLITERALconditions of this experiment represent these two possibilities,where the word is the same and the meaning is different. ThePARAPHRASE and NONLITERAL conditions hold meaning constant whileusing different words.The use of single word metaphors in both portions makescomparing the dead and novel metaphor portions more meaningfulthan in Experiment 1. Reading times for the targets should becomparable, so any response time differences between theNONLITERAL conditions in the two portions of this experiment canbe attributed to conventionalization differences between the twokinds of metaphors.To test whether all metaphors require essential processingthat goes beyond conventionalization, the novel metaphor portionstimuli of the earlier experiment were edited to select as muchas possible only those metaphors that were novel at least to thedegree that they were not listed as secondary definitions of theliteral word, based on Webster's New World Dictionary of AmericanEnglish (Neufeldt, ed., 1988). Conversely, dead metaphors wereselected if they were so listed. An example of a dead metaphoritem is provided in Figure 3.Metaphor Comprehension - 65CONTEXT Whenever Toby gets a new mystery novel by her favouriteauthor, she puts everything else aside because she finds itSOENDINGS [NONLITERAL] absorbing.[PARAPHRASE] interesting.[SURPRISE] difficult.LITERAL CONTEXTThe major oil companies are interested in developing amaterial for cleaning up oil spills that is harmless, yethighly oilENDING[LITERAL] absorbing.Figure 3: Sample dead metaphor item from Experiment 2.Metaphor Comprehension - 66In developing items for this study, it was interesting todiscover that most truly unique metaphors were not suitablebecause they were impossible to express in a single word for theparaphrase conditions. Metaphors that were too obscure were alsoexcluded, but for a different reason: it was felt they couldsimply be impossible for subjects to understand in a reasonablelength of time. As a result, Experiment 2 is a weaker test ofthe possible differences in processing time than mighttheoretically be possible if only absolutely unique metaphors hadbeen included. In effect, this experiment only compares highlyconventionalized metaphors to less conventionalized ones.MethodSubjects Thirty-two new undergraduate subjects were recruited fromthe University of British Columbia psychology student subjectpool. Each received one class credit for an hour ofparticipation. All claimed to be competent readers and speakersof English. Three subjects in this group were dropped andreplaced, one because a low post-test score (66%) combined withvery fast response times indicated that the subject was probablynot reading the stimuli to the point of understanding, and twoothers after a preliminary analysis, because their means in overhalf the conditions exceeded the group means by more than twostandard deviations.Stimuli and Procedure Stimuli in the NONLITERAL condition of this experiment wereMetaphor Comprehension - 67all chosen to be dead metaphors or novel metaphors. Thisessentially amounted to a partitioning of the metaphor conditionof Experiment 1, with some refinement and a number of new itemscreated. Twenty stimuli were developed for each portion, and thesame four conditions as in Experiment 1 were created for eachstimulus. Stimuli and targets used in this experiment and a copyof the post-test are provided in Appendix B.Dead metaphor and novel metaphor items were constructed inthe same way, with similar length contexts of one or twosentences. Contexts varied in length from 15 to 28 words in thecase of the dead metaphors, and 18 to 29 words for novelmetaphors. Targets, with a single exception, were all singlewords. The design, apparatus, and presentation used in thisexperiment were identical to those of Experiment 1. Everysubject received 160 trials of alternating blocks of 20 deadmetaphor items or 20 novel metaphor items in various orderswithin a single session.Results and DiscussionAll subjects obtained scores ranging from 78-100% on thepost-test, with the mean for this group falling at 93.6%. Thesuccessful retention of test items was apparently the result ofsubjects correctly following directions to read items to thepoint of understanding.Data were edited before analysis and any response timediffering by more than two standard deviations from theinaiviauai subject's mean in that condition was deleted, and theMetaphor Comprehension - 68mean recalculated based on the remaining response times. Anyresponse of less than 100 msec or greater than 5,000 msec wasalso excluded. This amounted to less than 5% of total responses.Analyses were performed both on the full range ofexperimental data and then just on the three base conditions, asexplained in Experiment 1. Order and even/odd ANOVAs arereported for only the three base conditions. The 2 X 4 repeatedmeasures analysis of variance showed statistically significanteffects for both portion [F(1,31) = 4.22, p = .048] and condition[F(1.4,42.7) = 27.24, p < .001] and the interaction between them[F (1.7,54) = 16.24, p < .001]. These results were substantiatedin the 2 X 3 analysis [portion: F(1,31)= 12.54, p = .001;condition: F(1.2,36.3)= 17.98, p < .001] and a significantportion by condition interaction [F(1.3,40.2) = 12.82, p < .001].Results are shown in Figure 4.The dead metaphor and novel metaphor portions show similarpatterns of response times except for the NONLITERAL condition,where the two metaphor conditions differed in familiarity. Theseresults are important in that most research on metaphor seeks toidentify differences between literal and nonliteral languagewithout clearly indicating any criteria for selecting nonliteralsamples, yet there is a dramatic difference between dead andnovel metaphors that would necessarily influence the conclusionsthat can be drawn. The dead metaphors used in this experimentproduced response times similar to literal words having the sameittaning (NONLITERAL [deadj, 765 msec; PARAPHRASE, 780 msec),Metaphor Comprehension - 69MEANS - EXPERIMENT 2110010009000A.COCC 800700D NOVEL111 DEADN^P^L^SCONDITIONS - EXPERIMENT 2Figure 4: Results of Experiment 2 - Dead and novel metaphor portions. (N)nonliteral [dead/novel metaphor] (P) paraphrase (L) literal (S) surprise.Metaphor Comprehension - 70while the novel portion shows that novel metaphors required aboutas much processing time as literal words whose meanings weresurprising (NONLITERAL [novel], 936 msec; SURPRISE, 940 msec).The final important contrast to observe is between theNONLITERAL condition results in the two portions. Mean responsetime to NONLITERAL [dead] was 765 msec and to NONLITERAL [novel]was 936 msec, a difference of 171 msec. By way of contrast, thetwo PARAPHRASE results differ by only 25 msec and the two LITERALresults are only 7 msec apart. The statistically significantresult for portion can be accounted for by the difference infamiliarity between the two types of metaphor. These resultsdisconfirm the idea that a special cognitive process is involvedin metaphor processing, and point in the direction oflexicalization as the most important factor contributing toresponse time delays in nonliteral language understanding.When the stimuli are broken down into three groups toinvestigate the effects of previous exposure to context, therewas no order by portion by condition interaction [F(1.8,54.9) =.77, p = .453], indicating that priming is not having adifferential effect on response times in the various conditions.Stimuli divided into arbitrary groups according to contextnumber were analyzed by comparing the results of even to oddnumbered contexts. No statistically significant interaction ofportion by condition by context number group was found [F(2,62) =2.25, p = .114] in this analysis.Given these results, it appears that at least two distinctMetaphor Comprehension - 71factors could be affecting response times. The first factor isthe degree to which the context is permitting the subject toanticipate a target meaning, as revealed by comparing thePARAPHRASE to the SURPRISE conditions. The fact that theanticipation should be the same until the point where the targetis displayed indicates that it is at least partially a quality ofthe target that determines how quickly it is understood. Thisquality is not simply determined by whether a word is literal ornot, since these two conditions both have literal endings. Itmust be the compatibility of meaning between context and targetthat is critical. In the present results, this advantageoperates in favour of the PARAPHRASE targets.An equally important result to emerge from the presentexperiment is that novel metaphors take longer than theirparaphrases to process. From this evidence, it is reasonable tosuspect that some kind of additional processing is going on whena word with a previously lexicalized literal meaning is usedmetaphorically in a novel way. However, the lack of responsetime difference between the dead metaphors and their paraphrases,suggests that any conclusions that could be drawn about metaphorprocessing would be inadequate without defining exactly howconventionalized all the metaphors being investigated might be.It was felt that it would be important to do an experiment basedon some objective criteria for conventionalization, both forliteral and nonliteral words.Metaphor Comprehension - 72EXPERIMENT 3 - THE EFFECTS OF CONVENTIONALIZATIONExperiment 2 provided evidence concerning the important rolethat conventionalization plays in the comprehension of metaphors.While it revealed that subjects respond differently to dead andnovel metaphors as a class, controls for the conventionalizationof any of the other targets were not imposed. As in previouslypublished experiments investigating literal/nonliteraldifferences, randomness with regard to degree ofconventionalization was assumed, but if there were acoincidentally higher overall level of conventionalization in thetargets of one condition than in another, such influences couldproduce spurious effects that would seriously affect theempirical findings. A case in point might be if the PARAPHRASEwords were more highly conventionalized as a class than theSURPRISE words. If so, the SURPRISE condition results could beslower for this reason alone, rather than because they disruptedthe anticipation generated by the context. Since all of thepreviously drawn conclusions could be challenged by thispossibility, Experiment 3 was designed to specifically addressthe question of conventionalization.If we define the mental lexicon as the words and wordmeanings any individual has access to as a result of his/herpersonal experiences with a language, it is virtually impossibleto obtain information about exactly how conventionalized anyMetaphor Comprehension - 73specific words are for any individual subject. Ideally, thisinformation would be necessary to support the claim that metaphorprocessing is determined by whether the metaphorical use of aword has become lexicalized. A practical alternative is toemploy a list of count values derived from word frequencyanalyses. Just as any word frequency count reflects the vagariesof the particular items chosen, any individual's mental lexiconreflects the vagaries of her/his own particular linguisticexposure. Yet word frequency lists are useful in a general sensebecause they reveal overriding frequency trends that exist in alanguage, particularly at the extremes, and these can beconsidered to parallel the linguistic exposure received bylanguage users. For this experiment, word frequency counts wereused as an objective lexicalization criterion.Experiment 3 used some of the stimuli from Experiment 2 in asimilar design but with the addition of an UNFAMILIAR conditionin both the dead and novel metaphor portions. Several new itemshad to be added in order to accommodate all of the endings. TheUNFAMILIAR condition consisted of the same context that was usedin the items in each of the other appropriate conditions (recallthat the LITERAL condition used a different context) but endedwith a literal word that was relatively unfamiliar, as determinedby its frequency of use in everyday English according to Francisand Kucera's Frequency Analyses of English Usage: Lexicon andGrammar (1982). Selection of target words was aided by theLongman Synonym Dictionary (1986) and Webster's New WorldMetaphor Comprehension - 74Dictionary (Neufeldt [ed.], 1988). Care was taken to selectunfamiliar words that were not so unlikely that they would not berecognized or understood by an undergraduate subject; the resultscould obviously have been manipulated by choosing totally obscureor archaic unfamiliar words but this was not the intent of theexperiment, and such alternatives were judiciously avoided.Targets in the other conditions were refined so that they wouldbe as equal in word frequency as possible within a stimulus item,and distinctly higher in frequency than the UNFAMILIAR targets.There is a history of empirical evidence in support of the'word frequency effect' (e.g., Rubenstein, Garfield & Millikan,1970). These experiments have all shown that subjects reliablyrespond faster to high frequency words than low frequency words.The justification for adding the UNFAMILIAR condition was to seeif similar word frequency differences would emerge in thisparadigm. For example, if differences in response time betweenthe UNFAMILIAR and PARAPHRASE conditions were found to be largerthan differences between the NONLITERAL and PARAPHRASEconditions, then attributing any processing time advantage to themere fact of a word being 'literal' becomes increasinglyproblematic. Further, if latencies in the above comparisons aresimilar, then it is worth considering that similar factors mightbe affecting them. Such a finding would cast further doubt onthe existence of a distinct metaphorical processing stage inlexical comprehension.Metaphor Comprehension - 75MethodSubjects A new group of 40 undergraduates from the University ofBritish Columbia psychology student subject pool served assubjects in this experiment. Each received one hour of classcredit for participation. All claimed to be competent readersand speakers of English. Two of the original subjects in thisgroup were dropped and replaced, one because of a high proportionof mean response times that exceeded the combined subject meansby more than two standard deviations, and one because of a scoreof 50% on the post-test.Stimuli and ProcedureThe stimuli in this experiment were of the same type as usedin the previous experiments, with the addition of the UNFAMILIARcondition. While a precise matching on word frequency for all ofthe other targets was obviously impossible, overall frequencycounts were adjusted so that they were similar for the fourreplicated conditions, and much lower for the UNFAMILIARcondition within each stimulus item. Average word frequencycounts in the dead metaphor portion as obtained from Francis andKucera (1982) were: NONLITERAL, 57.75; PARAPHRASE, 54.6; LITERAL,57.75, SURPRISE, 54.3; UNFAMILIAR, 1.35. In the metaphorportion, they were: NONLITERAL, 30.5; PARAPHRASE, 34.4; LITERAL,30.5; SURPRISE, 33.9; UNFAMILIAR, 1.7. To recapitulate, theNONLITERAL, PARAPHRASE, SURPRISE and UNFAMILIAR conditions alln en ica con ext. The targets of theMetaphor Comprehension - 76NONLITERAL, PARAPHRASE and UNFAMILIAR conditions also all hadmeanings that were similar, but the target for the NONLITERALcondition was a metaphor (dead or novel), for the PARAPHRASEcondition it was a literal word that was used with approximatelyequal frequency to the nonliteral word (as used in its literalsense), and for the UNFAMILIAR condition it was a literal wordthat was used much less frequently in normal discourse. TheSURPRISE condition presented the same context, but its targetshad a different meaning than the others, and one that was lesslikely to be anticipated, although its usage was of approximatelyequal frequency on the basis of the Francis and Kucera (1982)data. Very low frequency surprise words were not used, in aneffort to keep the UNFAMILIAR and SURPRISE categories as distinctas possible. As before, the LITERAL condition had differentcontexts but the same target word as the NONLITERAL condition,but with the target word used literally. Consequently, wordfrequency counts in these two conditions were always identical.An example of a novel metaphor item from this experiment is shownin Figure 5. A complete list of stimuli and targets, along witha sample of the post-test, appears in Appendix C.Some of the stimulus items from the previous experimentssurvived intact into the present one, but many were eliminated orrefined in order to bring them into line with the strictercriterion of similar word usage frequency. Stimulus setsconsisted of twenty dead metaphor and twenty novel metaphoreach having five variations, for a total of 100 of each,Metaphor Comprehension - 77CONTEXTThe politician's presentation was smooth, but the facts hepresented were not consistent with what Terry knew to betrue. She suspected the data had beenENDINGS [NONLITERAL] cooked.[PARAPHRASE] changed.[SURPRISE] stolen.[UNFAMILIAR] contrived.LITERAL CONTEXTMarilyn spent the morning in the orchard, picking applesfor applesauce. She didn't stop working until evening, whenthe apples were finally peeled andENDING[LITERAL] cooked.Figure 5: Sample novel metaphor item from Experiment 3.Metaphor Comprehension - 78or 200 presentations per subject.Sets were combined as before, with the additional conditionmaking each set of 20 consist of 4 samples of each of the fiveconditions. Dead metaphor and novel metaphor sets werealternated, with each subject receiving a unique order ofpresentation of the sets. As before, a 32 item, true/falsepencil-and-paper recall quiz followed the response time portionof the experiment. Presentation of stimuli, response procedure,and collection and analysis of data were all performedidentically to the method described in Experiment 1.Results and DiscussionPost-test scores in this condition ranged from 81-100%correct, with a mean of 93.3%. The high retention scoresindicate that items were being read to the point of understandingby all subjects.Responses differing by more than 2 standard deviations froman individual subject's mean for that condition were deleted,and the mean recalculated based on remaining responses. Thisresulted in the loss of about 5.5% of the total data. Theresults of this Experiment are displayed in Figure 6.When the results of the full experiment were subjected toanalysis using a 2 X 5 repeated measures ANOVA, both portion andcondition were revealed to exert statistically significanteffects [portion: F(1,39) = 7.87, p = .008; condition:F(1.5,57.6) = 37.16, p < .001]. The interaction of portion andcondition was not significant [F(2.5,97.4) = 2.28, p = .09].Metaphor Comprehension - 79MEANS - EXPERIMENT 31400C) 12002rz, 10000n.cc 800600D NOVELBI DEADN P^L^S^UCONDITIONS - EXPERIMENT 3Figure 6: Results of Experiment 3 - Dead and novel metaphor portionsmatched for word frequency. (N) nonliteral fdead/novell (P) paraphrase(L) literal (S) surprise (U) unfamiliar.Metaphor Comprehension - 80Focusing on just the three base conditions, the 2 X 3 repeatedmeasures ANOVA revealed a statistically significant differencefor portion [F(1,39) = 6.91, p = .012] and for condition[F(1.4,55.4) = 11.74, p < .001], and a statistically significantportion by condition interaction [F(1.5,57.9) = 6.56, p = .006].The interaction in these results is attributable to the slowerresponse times for NONLITERAL stimuli in the novel portion andthe faster response times for the LITERAL stimuli in the deadportion.To test for effects of priming based on presentation order,four groups based on order were formed and response times for thefirst vs. last presentations of a context were compared. Theportion by condition by order interaction [F(1.8,68.4) = .73, p =.470] was not statistically significant. These results may beregarded as evidence against any differential priming caused byrepeated exposure to the same context.Even vs. odd numbered stimuli for the NONLITERAL,PARAPHRASE, and LITERAL conditions were also subjected toanalysis of variance, to provide an indication of whether theseresults could be generalized to other stimuli. There was nointeraction of this factor with the portion by conditioninteraction [F(2,77.1) = 2.37, p = .101].This experiment indicates that subjects' previouslyestablished familiarity with the target words has a dramaticeffect on their response time to understand the target. Itshould be remembered that the UNFAMILIAR condition targets wereMetaphor Comprehension - 81all literal words having (approximately) the same meaning andpresented following the same context as both the NONLITERAL andPARAPHRASE conditions, so these three conditions are directlycomparable. There was no significant difference between theresponse times for the dead metaphors in the NONLITERAL condition(mean = 843 msec) and their literal counterparts in thePARAPHRASE condition (mean = 848 msec). The NONLITERAL novelmetaphors (mean = 922 msec) took an average of 95 msec longer tounderstand than the PARAPHRASE targets (mean = 827 msec) andaccount for the statistically significant portion by conditioninteraction in the 2 X 3 analysis. UNFAMILIAR targets, however,showed response time means of 1165 msec in the dead metaphorportion and 1214 msec in the novel metaphor portion. The factthat this difference overshadows the effect of the metaphoricalword suggests that what we define as nonliteral is beingconfounded with familiarity in response time experiments. Thatis, if we depend on response time differences to distinguishliteral from nonliteral usages without controlling for howfamiliar the words are, the differences we find might also beaccounted for in terms of the unfamiliarity of the metaphoricalusage rather than any distinct metaphorical quality. The factthat both NONLITERAL conditions are faster than theircorresponding UNFAMILIAR conditions suggests that metaphor mayhave a facilitory function in understanding. The argument couldbe made that metaphors are functional in this respect, and thewi n^nly voiced observation that it is often new information thatMetaphor Comprehension - 82is conveyed through metaphor is given additional credence bythese results.The SURPRISE condition results are also important. Becausethis condition has the same context as the PARAPHRASE conditionand is also completed by a literal word, a relatively puremeasure of the effect of the compatibility of context with targetword is possible when their respective response times arecompared, now that word frequency has been controlled. ThePARAPHRASE condition can be regarded as a baseline, since itstarget is a common literal word that fulfils the conventionalexpectations generated by the context. It should give thefastest response times of any of the conditions sharing the samecontext. According to the data, the effect of thwartingcontextual expectations in the SURPRISE condition slowsresponding twice as much on average as having a novelmetaphorical target does: the difference between the SURPRISE andPARAPHRASE response times was 192 msec, while the differencebetween NONLITERAL and PARAPHRASE response times was 95 msec.The difference between the SURPRISE and PARAPHRASEconditions shows that response times can be different despitehaving identical contexts and literal word endings of equalfrequency. This difference may be attributed to context-targetcompatibility. A word that is highly predictable within thecontext is easily understood, while a word that is lesspredictable, though no less grammatical, tends to disruptunderstanding to a marxed aegree. Since literal words have shownMetaphor Comprehension - 83response time differences that are both shorter (for paraphrases)and longer (for unfamiliar words) than those for nonliteralwords, it is not tenable to hypothesize an additional processingstage to account for the understanding of metaphors, at leastuntil we can isolate the effects of familiarity.On the other hand, the entire responsibility for determiningrapid understanding cannot simply be attributed to a compatiblecontext. The PARAPHRASE and UNFAMILIAR conditions had the samecontext and the same meaning, and yet produced different responsetimes. If context were the dominant source of meaning, neithermetaphors nor unusual or unfamiliar words should show anyresponse time differences. The slower times for the UNFAMILIARwhen compared to the PARAPHRASE conditions make it plain thatunderstanding is at least partly determined by the familiarity ofthe target word itself. Nor is it correct to say that ametaphorical target categorically takes longer to understand thana literal one, since these results show the opposite when theliteral target is surprising or unfamiliar.Even stronger evidence for the contribution of familiarityis provided by contrasting the results of the SURPRISE andUNFAMILIAR conditions. Both have identical contexts and literalword endings. The difference between them is that the contextsupports the meaning of the UNFAMILIAR targets but not theSURPRISE targets, and this lack of contextual support has beenshown in the two previous experiments to cause response slowing.By any 'compatible context plus literal word' hypothesis thatMetaphor Comprehension - 84ignores familiarity, the UNFAMILIAR targets should be faster thanthe SURPRISE targets, and yet they are not. The fact that theSURPRISE endings have higher word frequency counts than theUNFAMILIAR endings accounts for the difference, regardless of theUNFAMILIAR condition's contextual compatibility advantage.Metaphor Comprehension - 85EXPERIMENT 4 - VOLITIONAL METAPHOR GENERATIONIf dead metaphors can be understood as quickly as literalwords, can they also be produced just as readily? Is there adifference in the capacity to generate novel as opposed to deadmetaphors? This experiment looks at metaphor from the standpointof volitional generation in response to a context. The evidenceof the three previous experiments suggests that novel metaphorsare unlexicalized uses of lexicalized words. It seems unlikelythat we would have independent mental lexicons for comprehensionand for generation, but if we generate speech just from the wordswhose meanings we have lexicalized, where do metaphors (or moreprecisely, novel metaphors) come from? The implication from theprevious experiments in this thesis is that novel metaphorgeneration should come only from a linguistic need that is notrepresented by a previously lexicalized word. Since the wordchosen has to have enough common aspects that the new meaning canbe made comprehensible through its use, it would seem that such aselection would need to be carefully made, and should take longerthan generating highly lexicalized words.Schank's (1982), Glucksberg's (1991), and Chandler's (1991)models all explain metaphor comprehension by an expansion of themeaning of the vehicle to incorporate the topic. Cutting acrosstheories somewhat, we could say that in generating a novel y a wor•as a thematic relationship that-^• •^-^• - n 1Metaphor Comprehension - 86could be applied to another concept, and label the theme with theconcept prototype from which it arose. This would mean that themetaphorical meaning did not exist unless it was needed.Attention in this fourth experiment is directed towardilluminating the question of the lexicalization of metaphors.When subjects are given the opportunity to freely generate anykind of sentence endings, do they respond by accessinglexicalized words first? If so, would dead metaphors be amongthose so accessed? Do the endings they can generate parallelword frequency information - i.e., are the first endings subjectsgenerate high frequency words, with lower frequency words andnovel metaphors coming afterwards? Is it even possible togenerate novel metaphors on demand?In this experiment, subjects were asked to generate theirown single word endings for some of the same sentence contexts aswere used in the previous experiments to see what kinds ofendings would be generated, how consistent they would be betweensubjects, whether the order in which words were generated wouldcorrespond to printed word frequency data (as an indication oflexicalization), and whether novel metaphors would occur at allamongst the generated data.MethodSubjects Sixteen undergraduate psychology students from theUniversity of British Columbia participated in the experiment forone hour of class credit. These students had not participated inMetaphor Comprehension - 87any previous experiment in this series.Stimuli and Procedure Subjects were given an instruction sheet to read and thenwere handed a booklet consisting of 20 brief (1-2 sentence)contexts, in which the last word of the last sentence wasmissing. Three contexts were presented per page, and subjectswere told to work on one until a buzzer sounded as a signal tomove on to the next. Two minutes were allowed for each contextitem. A countdown was displayed on a video screen beside thesubject, to help them keep track of what context number they wereon.The contexts were a selected group of stimulus items fromExperiment 3, presented verbatim without targets. Each contextwas followed by a total of 12 blank lines. The first 10 lineswere intended for subjects to write as many suitable endings tothe context as occurred to them, and the last two were forendings that might suggest themselves while working on anotherquestion, after the time interval was over. Subjects werecautioned not to fill up spaces with synonyms of a single idea,and flexibility of approach, including the possibility ofmetaphors, was specifically encouraged. Responses were screenedfor appropriateness, grammaticality and metaphoricity by theauthor and by an independent coder, who then collaborated toresolve discrepancies. All contexts and tabulated responsesappear in Appendix D.Metaphor Comprehension - 88Results and DiscussionSubjects were able to come up with a surprising number ofplausible possibilities in response to the stimuli, averagingmore than 6 per context per subject. Further, there was littleintersubject overlap, resulting in a large total number ofdifferent responses. Overall, 79% of raw responses were uniqueto a single subject. Even after editing to remove ungrammaticalresponses, the minimum number of responses generated by the 16subjects in any context was 57 and the maximum was 162.The nature of the contexts was such that most responses wereadjectives describing some aspect of a person or situationmentioned in the contexts. The overwhelming majority ofresponses (92%) were words used literally. The definition ofliteral word usage was based on the intuitive judgement of thetwo coders, who proved to have been highly consistent. Virtuallyall of the remaining responses that could be classified asmetaphors (8%) were the most conventionalized kind of deadmetaphor (e.g., brilliant, cold, distant). Figure 7 illustratesa typical context taken from the experiment and the number andvariety of words produced in response to it.Within each context, words that occurred a single subjecthad a significantly lower printed word frequency (Kucera andFrancis, 1968) than did words that occurred to more than onesubject [F(1,19) = 6.73, p = .018]. The data from this analysisare illustrated in Figure 8.The first word that occurred to a subject in a specificMetaphor Comprehension - 89Figure 7: Sample item from Experiment 4, with responses.STIMULUS CONTEXT AND EDITED RESPONSES Angry parents wanted the government to stop cutbacks to schools, but theDepartment of Education maintained that their plans wereresponse [#of Ss]^** = metaphor1. acceptable [1]2. adequate [2]3. appropriate [1]4. approved [1]5. bad[1]6. [the] best [1]7. changing [1]8. continuing [1]9. correct [4]10. critical [1]11. decided [2]12. definite [1]13. demanded [1]14. different [1]15. established [1]16. excellent [1]17. fair [1]18. final [2]19. finalized [1]20. finished [1]21. firm [2]**22. fixed [1]23. foolproof [1]24. good [2]25. great [1]26. ignored [1]28. implemented [1]29. important [1]30. indifferent [1]31. inevitable [1]32. inflexible [1]**33. intelligent [1]34. irreversible [1]35. justifiable [1]36. justified [1]37. law [1]38. long-range [1]39. made [2]40. minimal [1]41. necessary [3]42. normal [1]43. official [1]44. overdue [1]45. permanent [1]46. positive [1]47. practical [1]48. proceeding [1]49. progressing [1]50. proven [1]51. reasonable [3]52. responsible [1]53. responsive [1]55. righteous [1]56. rigid [1]**57. settled [1]**58. silly [1]59. substantial [1]60. superior [1]61. supportive [1]62. transferred [1]63. unchangeable [2]64. unstoppable [1]65. untransferable [1]66. useful [1]67. useless [1]68. wanted [1]69. well-considered[1]70. wise [1]71. workable [1]72. working [1]27. immutable [1]^54. right [3]Metaphor Comprehension - 90context was also more likely to occur to other subjects than thelast word was [F(1,19) = 28.65, p = .001]. These data are shownin Figure 9. On the other hand, there was no indication thatthese first words had a higher printed frequency than the lastwords [F(1,19) = .63, p = .438].These results support the belief that novel metaphorsrepresent unlexicalized uses of previously lexicalized words.The absence of any but the most conventionalized metaphorssuggests that producing novel metaphors is very likely notachieved through direct lexical access. It appears that a taskof this type encourages subjects to search their personal lexiconfor suitable word endings, and novel metaphors are not present inthat lexicon.Individually, the first completion that occurs to mostsubjects is often the same. Later, the possible ways ofconstruing the context tend to diverge, but the frequency ratingsof the words selected does not decline. This suggests that thelexicon is not organized as a single dictionary, but rather thatit may be parsed into discrete units based on semantic similarity(Jastrzembski, 1981). Subjects appear to be moving from unit tounit rather than scanning an entire list, whether orderedalphabetically or by overall frequency, although frequencyinformation may play a part within a unit.These results are also revealing in terms of the contribution ofcontext to understanding. None of the contexts was tightlyconstraining to subjects, despite the fact that only the lastMetaphor Comprehension - 9110090>-0z800LLa 70cc06025040ONE^ONESUBJECTS MAKING SAME RESPONSEFigure 8: Results of Experiment 4 - Comparison of the printed wordfrequency of endings generated by only one subject in a context with thenumber generated by more than one subject.Metaphor Comprehension - 926Luz0a. 5cc2 4U)0z32z401 FIRST^LASTORDER OF OCCURENCE OF RESPONSEFigure 9: Results of Experiment 4 - Comparison of mean number  of subjects making the same response for first generated and last generatedwords.Metaphor Comprehension - 93word of the last sentence was missing. While the finding thatfirst words were more likely to be consistent between subjectsthan last words were suggests that there is likely someculturally controlled tendency to construe contexts similarly,the imbalance of different word endings over repeated onestestifies to an amazing flexibility and openness in language thatfurther argues against understanding dominated by contextualconstraint.Metaphor Comprehension - 94EXPERIMENT 5 - SEMANTIC DECISIONResponse latency to single word endings proved useful as ameasure of comprehension for the first three experiments,precisely because it produced the confounding of familiarity withmetaphoricity that led to attempts to isolate these factors inthe later experiments. At the same time, however, it presented aserious problem. Since the presentation subjects were givenconsisted of the context followed by a single word target andrequired only a keypress when the target was understood, subjectscould hypothetically keypress their way through the entireexperiment although their understanding of the targets might bevariable, vague, or nonexistent. The obtained results could be aconsequence of such a tactic, since longer hesitations would bemore likely in the case of the more unusual targets in theNONLITERAL, SURPRISE, and UNFAMILIAR conditions regardless ofwhether they were ultimately understood. Even the post-testmight not ensure that the results were valid, as many questionscould be answered by recalling the meaning of any of the variouswords that completed the context. Because of these possiblesources of error, and despite the fact there has been a well-established precedent for using simple response time inexperiments of this type (Hoffman and Kemper, 1987), it was feltthat the conclusions that could be drawn from these results wouldbe more valid if they could be replicated using a differentMetaphor Comprehension - 95paradigm.Experiment 5 was intended to replicate Experiment 3, usingchoice response time to a semantic decision task. Hoffman andKemper (1987) have pointed out that time to understand metaphorsmay not be a reliable indication of what is actually going onwhen experimental subjects are presented with artificiallycreated metaphors. Nonetheless, in any well designed experiment,consistent differences must offer evidence of some psychologicalreality, and the refinements produced by employing choiceresponse time potentially provide more information than simpleresponse time does. A further advantage of the choice task isthat it offers some insight into where and how the delays thatinfluence response time arise. Specifically, the data from thisexperiment were analyzed by counting the number of errors in eachcondition, and separating these from the responses droppedbecause of response time violations (i.e., longer than twostandard deviations beyond the individual subject's mean in eachcondition or less than 100 msec or more than 10,000 msec).Doing this provides an additional perspective on why someconditions take longer than others. For example, UNFAMILIARwords might be expected to produce a bipolarity of responsedepending on whether a target is recognized at all. This wouldindicate that if a word is known, it presents no problem, but ifit is not known, subjects are simply unable to make use of it,and must resort to what they can extract from context, resultingin a cluster of longer times on some of the items. An errorMetaphor Comprehension - 96analysis should also prove enlightening about what happens whenmetaphors are encountered. In Experiment 3, the SURPRISE andUNFAMILIAR words took longer in both portions than NONLITERALwords. Does this mean that the metaphors are more efficientlyconveying their meaning, or is it that they are perceived as soincongruous that the subject considers them nonsense, and simplyresponds randomly? Discovering whether subjects make more errorsin one condition than another would help answer this question.MethodSubjects Twenty new subjects completed this experiment. Most wereundergraduate students at the University of British Columbia whoparticipated for class credit in psychology, while the remainderwere paid volunteers. Six of the original subjects were droppedand replaced due to combined error and response delay deletionsexceeding 25/200, or 12.5% of total responses.In addition, a pre-experiment was run to validate the choiceof target and non-target items. Five subjects completed thisportion, but two clearly misunderstood a large proportion of theitems as well as the task. These respondents gave evidence thatEnglish was not their first language, and their results were nottaken into consideration as validation data.Stimuli and ProcedureIn this experiment, a semantic decision task was employedwhere subjects had to make a choice between an appropriate and aninappropriate target with respect to meaning. Subjects wereMetaphor Comprehension - 97presented with a 1 - 2 sentence context, displayed all at once ona video screen. Once they felt they had read and understood thecontext, they were to press the bottom buttons on the responsebox. This removed the context and presented an appropriate andan inappropriate target word, side by side, on the screen. Theywere to choose the word that made the most sense in terms of thecontext they had just read, and to indicate their choice bypressing a button on the right or left, corresponding to theposition on the screen of the correct target word.Stimulus contexts were taken from Experiment 3. Each targetwas paired with a nontarget word of identical length and closelymatched printed word frequency count, as determined from Kuceraand Francis (1968). Nontargets were also selected forgrammatical appropriateness to the sentence context. All stimuliand targets used in this experiment are included in Appendix E,along with a copy of the post-test.A pre-experiment was undertaken in which three subjects wererun through the entire experimental procedure, but rather thanresponding as quickly as possible to their choice of a targetword, they were asked to rate each word, target and nontarget, ona seven point scale for comprehensibility and appropriateness;that is, whether each word made sense, given the context. Arating of 1 indicated 'highly appropriate' and a rating of 7indicated 'highly inappropriate'. Any items for which a targetwas given a rating of 4 or more, or a nontarget was given arating of 4 or less, or the two words were given ratings lessMetaphor Comprehension - 98than 4 scale values apart, were identified and replaced in orderto establish a clear distinction between targets before the mainpart of the experiment was run.For the actual experiment, the presentation of sentencetargets proceeded as in Experiment 3, using identical equipment,except that two words appeared on the screen side by sidefollowing the removal of the context, rather than just one.Subjects used a four button response box with the thumbs of bothhands pressing the two bottom buttons simultaneously to presentthe target word choices after the context had been read andunderstood, and the index fingers of each hand pressing the topbutton on the right or left corresponding to the choice of thecorrect target. This posture was designed to standardizeresponse time, eliminating the variability introduced by subjectsmoving their hands onto the button before pressing it. The two-thumb technique was also intended to prevent a hand bias frominfluencing the target response. Choice of side for occurrenceof each correct target was randomly selected for each subjectwithin the presentation program. Choosing a target wordautomatically resulted in the presentation of the next stimuluscontext in the set.Stimuli were presented in sets of 20, with the order ofpresentation of sets randomized between subjects, except thatdead metaphor and novel metaphor sets always alternated, just asin Experiment 3. Following the stimulus presentation portion ofthe experiment, subjects were given a 32-item true/false pencil-Metaphor Comprehension - 99and-paper quiz, as a means of ensuring that they had actuallyread and understood the stimulus items.Results and DiscussionScores on the post-test ranged from 81-100% correct, with amean of 93.5%. These results strongly indicated that subjectsexperienced no difficulty in correctly reading or understandingthe stimuli.Data were screened to eliminate incorrect responses and anytimes under 100 msec or over 10,000 msec. Based on the remainingdata, all response times greater than 2 standard deviations froman individual subject's mean for a given condition were alsodropped and the mean recalculated. Any subject for whom thisscreening procedure resulted in the elimination of more than 25responses (12.5% of their data) was dropped from the analysis andreplaced by a new subject. For subjects remaining in theanalysis, data dropped by all three screening processesrepresented 381 responses in total, or 9.5% of all data - 3.8%because of decision errors and 5.8% because of time delays.A portion by condition (2 X 5) analysis of variance ofresponse times for the whole experiment revealed main effects forboth portion [F(1,19) = 10.9, p = .004] and condition[F(2.1,40.7) = 49.37, p < .001], as well as a statisticallysignificant interaction between the two factors [F(2.4,45.8) =11.86, p < .001]. When only the NONLITERAL, PARAPHRASE, andLITERAL conditions were analyzed, the results were consistentwith the five condition analysis (portion: [F(1,19) = 13.42, p =Metaphor Comprehension - 100.002]; condition: [F(1.3,24.1) = 26.1, p < .001]; interaction:[F(1.2,22.8) = 21.72, p < .001]). Results are shown in Figure10. Clearly, the portion by condition interaction arose from thedifference in response times between the two portions in theNONLITERAL condition.The order in which a context was viewed (first vs. lastexposure) did not interact significantly with the portion bycondition interaction [F(1.3,25.2) = 1.8, p = .192]. Whenstimuli were divided into two arbitrary even/odd numbered halvesand compared as a measure of generalizability, the half byportion by condition interaction was not statisticallysignificant [F(2,38) = 2.14, p = .132]. These analyses may beinterpreted to indicate that priming effects are not responsiblefor the portion by condition interaction, and that thecharacteristics of the stimuli are not idiosyncratic.In reaction time experiments, Posner (1978) hasdistinguished the faster times recorded in simple as opposed tochoice reaction time as the difference between automaticactivation and conscious attention. This was not intended to bethe case in these studies, as any understanding task necessarilyrequires conscious attention, and reading the preceding contextdecisively separated Experiment 3 from studies where respondingwas based simply on detection. However, the necessity of readingtwo words on the screen as opposed to one, and of actually makinga decision as to which was correct, was expected to lengthenresponse times in this experiment in contrast to those of20001000Metaphor Comprehension - 101MEANS - EXPERIMENT 5N^P^L^S^UCONDITIONS - EXPERIMENT 5D NOVEL■ DEADFigure 10: Results of Experiment 5 - Dead and novel portions in a semantic •^I I I I I !^!^!^' 'I •(S)surprise (U)unfamiliar.Metaphor Comprehension - 102Experiment 3. This effect occurred as expected, but the trendsof the two experiments, with one notable exception, were much thesame. This congruency lends support to the validity ofconclusions drawn from the earlier experiments.The most notable change in the pattern of results of thepresent experiment from that of Experiment 3 occurred in theSURPRISE condition, where response times were relatively longerthan in the previous experiments. This condition also producedsubstantially more errors than any of the other conditions.Counts of the total number of errors occurring in each conditionof Experiments 5 and 6 are provided in Table 2. Since the errorrates are assumed to represent cases where subjects haveinsufficient information to make a correct decision and areforced to guess, it is probably safe to assume that the number oftimes this happens is roughly twice the obtained error rate -that is, an equal number of times when the subject makes a guess,that guess is fortuitously correct.In a count of error by condition, the SURPRISE conditionshowed 30 errors in the dead portion and 57 in the novel. Inboth portions this was more than twice as many errors as occurredin all other conditions combined. The 2 X 5 analysis of variancefor errors showed a significant effect for portion [F(1,19) =10.72, p = .004] and condition [F(2.3,44.6) = 23.84, p < .001]and a significant interaction of portion and condition[F(2.7,52.2) = 5.35, p = .004], attributable largely to theSURPRISE condition results. At the same time, the number of timeMetaphor Comprehension - 103violations stayed fairly consistent over conditions.Given the above results, it seems that when contextualexpectations are not supported, subjects become confused. Withthe single word targets of Experiment 3, they can simply do amental 'double take' if necessary to make sense of what appearson the screen, and force a new meaning out of the context theyhad just read. This process takes time, and is probablysometimes incomplete or unsuccessful, but the single buttonresponse paradigm is insensitive to these failings, as subjectswill press the button after attempting to derive meaning from thepresentation regardless of how incomplete their understanding is.In the semantic decision design of this experiment, subjectsencountering an ending that is different than they anticipatedmay attempt to test both targets in succession both before andafter they revise their interpretation of the context. The higherror rate may reflect the tacit time pressures, where subjectsconfused by having their expectations thwarted may conclude thatneither target makes sense and simply respond with a randomguess. In contrast, there were far fewer errors in theUNFAMILIAR conditions, about the same number of time violations,and mean response times were faster by about 80 cosec. Theadvantage for subjects in this condition could be that themeaning of the context was 'dependable' so the task might besimply to use whatever lexicalized information they had availableabout either the target or the nontarget to guide them to acorrect choice more often. No 'double-take' would beTABLE 2.Metaphor ComprehensionERROR AND TIME VIOLATIONS: EXPERIMENTS^5 and 6- 104EXP 5^EXP 6 EXP 5 EXP 6PORTION CONDITION ERRORS TIMEDEAD NONLITERAL 8 12 24 25DEAD PARAPHRASE 6 7 25 26DEAD LITERAL 1 2 22 25DEAD SURPRISE 30 38 24 24DEAD UNFAMILIAR 14 11 24 26NOVEL NONLITERAL 12 51 25 21NOVEL PARAPHRASE 4 4 21 22NOVEL LITERAL 6 20 17 21NOVEL SURPRISE 57 33 27 23NOVEL UNFAMILIAR 13 24 21 22Metaphor Comprehension - 105advantageous, and the longer response times could simply bereflecting time to summon more incidental clues, such asmorphemic information. The combination of longer response timesand higher error rates in the SURPRISE conditions draws attentionto the critical role played by context in developing consistentexpectations.The suggestion that the expectations of context exert strongconstraints on understanding might be challenged by citing theresults of Experiment 4, where it was shown that contexts wereinterpreted in any number of different ways. However, it isimportant to remember that in Experiment 4 subjects wereexplicitly instructed to use the contexts flexibly, and not toparaphrase the same idea in different words. This might haverequired them to duplicate the 'double-take' reading of thecontext each time. In any event, the results of this experimentcan not be taken as weakening the conclusions to be drawn fromthat one, or vice versa.This result provides new information regarding theimportance of context in comprehension which the previousexperiments may have disguised. Its full impact, in light of theresults of Experiments 3 and 4, will be discussed later.Briefly, it can simply be said that any less anticipated ending,whether NONLITERAL, SURPRISE, or UNFAMILIAR, will produce somedelay, but the fact that the meaning expectations of theNONLITERAL and UNFAMILIAR conditions remain consistent helpsthese words to be understood more effectively than the SURPRISEMetaphor Comprehension - 106condition, which requires a revision of how the context isinterpreted to be understood.One of the most interesting findings of this experiment isthat understanding due to the presence of a metaphorical targetis clearly not as much of a handicap to achieving understandingas either having an unanticipated or an unfamiliar target. Inboth portions, the NONLITERAL conditions produced fewer errorsthan the UNFAMILIAR and SURPRISE conditions, and no more timeviolations. Further, 14 of 20 subjects made no errors at all onthe novel metaphors and 13 of 20 made no errors on the deadmetaphors.From this vantage point, it begins to look as ifmetaphorical usages of conventionalized words might actuallyprovide a cognitive advantage over either the SURPRISE orUNFAMILIAR conditions, by allowing the use of the more familiarmeaning of the metaphorical word to assist the hearer,particularly if the information to be conveyed is new(surprising) or incomprehensible (unfamiliar). To furtherinvestigate whether metaphor actually contributes tounderstanding, and to gain further insight into why the SURPRISEcondition of Experiment 5 produced such a high rate of errorsrelative to all of the other conditions, it seemed useful to seewhat might happen if context were made a less useful component ofunderstanding. If contextual information were reduced to a bareminimum, would the SURPRISE target simply become impossible tointerpret. Might the lack of well developed contextualMetaphor Comprehension - 107expectations hinder the comprehension of novel metaphors?Ortony et al.'s (1978) work with long and short contexts appearedto point in this direction. It was in pursuit of answers tothese questions that Experiment 6 was developed.Metaphor Comprehension - 108EXPERIMENT 6 - REDUCED CONTEXTExperiment 6 focuses attention on the role of context inunderstanding. In Experiment 5, the delay in understanding timecaused by a SURPRISE target exhibited itself in both longerresponse times and higher rates of errors than in any of theother conditions. The SURPRISE condition responses took evenlonger than the UNFAMILIAR condition responses, which is areversal of direction from the Experiment 3 results. Given thisturn, it now seemed appropriate to delve more deeply into thecomponent of contextual preparedness, both to verify theintuitions raised by the results of the previous experiment andto further explore their implications.The presentation for this experiment was identical toExperiment 5, but contexts were rewritten to greatly decreasetheir length and informational content. Subjects in the twoexperiments were matched by presentation order. It was expectedthat reduced contexts would still provide sufficient informationfor rapid processing of PARAPHRASE and LITERAL conditions, butthat the UNFAMILIAR and SURPRISE conditions would suffer - theformer because the context was less able to create a full meaningthat would allow a less familiar word to be recognized asinformative, and the latter because the lack of contextualsupport combined with decreased contextual information would makethe choice between targets particularly difficult to resolve.Metaphor Comprehension - 109Since the SURPRISE condition would be more confusing, errors wereexpected to be high; the UNFAMILIAR condition could just showlonger response times. Results in the NONLITERAL condition wereless predictable from prior results. If the novel metaphorsmaintained their advantage over the SURPRISE and UNFAMILIARconditions when context was reduced, this finding would supportthe hypothesis regarding the functional advantage of novelmetaphors put forth in Experiment 5. The NONLITERAL condition inthe dead metaphor portion was again predicted to produce resultssimilar to its PARAPHRASE counterpart. Decreasing contextualinformation would be expected to affect both the PARAPHRASE andLITERAL conditions similarly, by lengthening response time andincreasing errors, but to a lesser degree than in the otherconditions.The attempt to reduce the length of the stimulus contextshad parallels with Ortony et al.'s (1978) experiment wheremetaphor understanding was measured in long and short contextconditions. In their experiment, the short contexts consisted ofjust the first sentence or phrase of their 5 - 6 sentence longcontexts, followed by the target. The example offered in theirpaper, with short contexts underlined here, is as follows (p.467) :Literal inducing context:Approaching the enemy infantry, the men were worried abouttouching off landmines. They were very anxious that theirpresence would be detected prematurely. These fears werecompounded by the knowledge that they might be isolated fromtheir reinforcements. The outlook was grim.Metaphor Comprehension - 110Metaphorical inducing context:The children continued to annoy their babysitter. She toldthe little boys she would not tolerate any more badbehavior. Climbing all over the furniture was not allowed.She threatened to spank them if they continued to stomp,run, and scream around the room. The children knew that herspankings hurt.Target:Regardless of the danger, the troops,(sic) marched on.Given the example provided, one can see that the strategy oflimiting a presentation to just the short context plus targetdoes more than just reduce information. The reader could be leftwith virtual nonsense. In the literal example, the semanticconjunction of the words 'enemy infantry' in the context with'danger', and 'troops' in the target is likely enough to triggera response indicating that the target was understood, althougheven prolonged contemplation of the presentation would yieldlittle real understanding. For the metaphorical presentation, nosuch conjunction occurs and the lack of coherence is manifest.It is hardly surprising that understanding time was slowed whencontexts were shorter. The intention underlying the presentexperiment was to reduce context information, but to maintainenough relevant cues so that targets could be clearlydifferentiated.Rather than simply processing a context until enoughinformation is gleaned so that a metaphor makes sense, it seemsmore reasonable to claim that the reader is continually involved-rpre a ion process, invo ving e ements ofMetaphor Comprehension - 111both context and target. Finding that metaphors are not slowedwhen context is shortened and the quality degraded would castdoubt on Ortony et al.'s findings with regard to the relationshipbetween context length and metaphor, suggesting instead thatOrtony's et al.'s shortened contexts slowed metaphorunderstanding because they lacked sufficient comprehensibility,rather than that were simply shorter in length. The length ofcontexts in both experiments was roughly equal, averaging 6 wordsin Ortony et al's study and 6.8 words in the present one. Thealternative possibility is that when contextual support isreduced, the NONLITERAL novel metaphors would produce the samekind of confusion as the SURPRISE condition showed in Experiment5, resulting in the same consequences of longer response timesand higher error rates.MethodSubjects A new group of twenty subjects from the University ofBritish Columbia subject pool participated in this experiment forclass credit. All indicated a fluency in English. Two subjectswhose error rates alone exceeded 15% and whose post-test resultswere 75% or less were excluded from analysis. A third subjectwas dropped because of response latencies that exceeded the meansof all other subjects by two standard deviations in more thanhalf of the conditions, and another was eliminated when sherevealed that she had participated in one of the earlierexperiments in this series. Four replacement subjects providedMetaphor Comprehension - 112satisfactory data.Stimuli and Procedure The stimuli used in this experiment were all deriveddirectly from the stimuli of Experiment 5, and are presented inAppendix F, along with a copy of the post-test. They werewritten with the intention of reducing both the length andinformation content of the contexts as much as possible, whilepreserving enough of the gist to make the targets comprehensible.Most proper nouns were changed to pronouns, unless multiplepronouns made reference impossible to determine. Dead metaphorcontexts were reduced from a mean length of 23 words as theyappeared in Experiment 5 to between 3 and 12 words (mean = 6.58).Novel metaphor contexts were reduced from a mean length of 25words to between 3 and 10 words (mean = 7.08). With a few slightexceptions to preserve readability/grammaticality, all editedcontexts contained exactly the same words in the same order asthe originals. An example of an item from Experiment 6 is shownin Figure 11, along with its source item from Experiment 5 forcomparison. Targets to items in Experiment 6 were identical tothose of Experiment 5, as were nontargets, with a few exceptionswhere the reduced context made the nontarget newly plausible, andan alternative nontarget with a similar word frequency count wassubstituted. Apparatus and procedure were identical to thosedescribed for Experiment 5.The correspondence between the stimuli of Experiment 5 andthose of Experiment 6 provides an opportunity to compare theseMetaphor Comprehension - 113EXPERIMENT 5 CONTEXT Edmund had written two novels before he turned 30. Laterhe wrote three more, but his editors rejected them all.They complained that his work had becomeEXPERIMENT 6 CONTEXTThey complained his work had becomeENDINGS [NONLITERAL] stagnant tapestry.[PARAPHRASE] dull sand.[SURPRISE] childish tropical.[UNFAMILIAR] hackneyed laminated.EXPERIMENT 5 LITERAL CONTEXTIt was shocking for Arnie to revisit his old home town after20 years away. Ugly townhouses replaced open fields, andthe pond where he often swam was nowEXPERIMENT 6 LITERAL CONTEXTThe pond where he swam was nowENDING[LITERAL] stagnant perverse.Figure 11. Sample novel metaphor item from Experiment 5 and itscounterpart from Experiment 6.Metaphor Comprehension - 114experiments to see exactly how much difference the contextmanipulation made. Although the two experiments were run ondifferent groups of subjects, the order of presentation in bothwas matched between groups, and a number of between-groupsanalyses were conducted. The results and discussion in thissection is limited to Experiment 6. An evaluation of theimplications of the two experiments taken together will be foundin the following section.The Experiment 5 post-test was edited and items truncated toeliminate information no longer contained in the stimuli. Thismade them somewhat vague, and it was anticipated that scoringwould demand more leniency than in previous experiments.Results and DiscussionScores on the post-test ranged from 63-94% correct, with amean of 81.05%. Due to the vagueness of some of the questions,subjects scoring at the low end of the scale were retained iftheir data were otherwise acceptable. The overall analysis ofvariance for five conditions showed statistically significanteffects for portion [F(1,19) = 63.66, p < .001] and condition[F(3.3,62.2) = 44.81, p < .001], and a significant portion bycondition interaction [F(3.8,72.6) = 11.51, p < .001]. The threecondition ANOVA produced identical levels of significance tothose for the five conditions [portion: F(1,19) = 83.75, p <.001; condition: F(2,38) = 20.04, p < .001; interaction:F(1.5,28.3) = 22.66, p < .001]. The results of Experiment 6 areshown in rigure 13. Again, the NONLITERAL metaphors in the novel20001000O NOVEL111 DEAD1500Metaphor Comprehension - 115MEANS - EXPERIMENT 6N^P^L^S^UCONDITIONS - EXPERIMENT 6Figure 12: Results of Experiment 6 - Dead and novel portions in a semanticdecision task with reduced context. (N)nonliteral [dead/novel metaphor](P)paraphrase (L)literal (S)surprise (U)unfamiliar.Metaphor Comprehension - 116portion were responded to more slowly than their PARAPHRASEcounterparts, while the dead metaphor portion produced about thesame response latencies in both the NONLITERAL and PARAPHRASEconditions.Looking at the effects of first vs. last (fourth)presentation of the same stimulus, it was found that there was noorder by portion by condition interaction [F(1.6,31) = 1.29, p =.286], indicating that repeated presentations of the same contextwere not priming one condition more than another.The effect of dividing the stimuli into two groups based onwhether arbitrarily assigned item numbers (but having the sameassignment as the items in Experiment 5) were odd or evenproduced a statistically significant interaction of arbitraryhalf with portion and condition [F(1.6,30.8) = 8.87, p = .002].Such results, along with those of the order analysis, indirectlyargue in favor of the repeated measures design of the presentstudies, since priming effects were found to be negligible, whileitem differences and intersubject differences can besubstantial.An error count was made to determine if reducing context hadaffected one condition more than another. Each subject received20 presentations in each condition and 20 subjects were used,resulting in 400 presentations per condition across subjects, and4,000 total presentations for the 10 conditions.^Data discardedas the result of errors accounted for 5.0% and time violationsaccounted tor b.8, for a total deletion of approximately 10.9%Metaphor Comprehension - 117of all responses. The error and response time violation data areprovided in Table 2. Error results were examined in a 2 X 5portion by condition ANOVA, which revealed a statisticallysignificant effect for portion [F(1,19) = 42.56, p < .001] andcondition [F(3.4,64.9) = 18.33, p < .001] as well as asignificant interaction of the two [F(3.4,65.2) = 11.86, p <.001]. The number of errors in the novel portion was almosttwice the number in the dead portion. Post-hoc Tukey tests, withfamilywise error set at a = .05, showed that the NONLITERALcondition of the novel metaphor portion had significantly moreerrors than any other condition. Errors in this condition alonerepresented 25% of all errors for both portions combined. TheSURPRISE condition for both portions also produced more errorsand differed significantly from all but the novel portionNONLITERAL and UNFAMILIAR conditions.One completely unexpected result that emerged from thisexperiment was that, in the novel portion, response times for theLITERAL condition were now significantly slower than for thePARAPHRASE condition and error rates were correspondingly higher.The reverse had been the case in every previous experiment. Eventhat finding was somewhat curious, since both the LITERAL andPARAPHRASE conditions could be described similarly as presentinga literal word target following a compatible context. It wouldseem most likely that response times should be about equal forthe two conditions, but if not, the expectation might be in theother direction, since the PARAPHRASE contexts had appearedMetaphor Comprehension - 118previously with the same meaning in both the NONLITERAL andUNFAMILIAR conditions, while the LITERAL condition context hadnever been seen before. Any priming effects as a result ofprevious exposure should favor the PARAPHRASE responses.Two possible explanations for the findings of the earlierexperiments were either that it was the previous exposure to thetarget, rather than the context, that was responsible forpriming, or that repeated exposure to the same context actuallyhad an inhibiting effect, even when understanding wasstraightforward, since the context was not always 'reliable'because of the SURPRISE condition. The former explanation seemsdubious, given that the target word was required to have twodifferent interpretations in the two exposures. The latter ismore plausible, although order analyses consistently demonstratedthat there was no statistically significant difference inresponse time between first and last presentations of contexts.It is only possible to speculate, but one explanation forthe earlier results is that, if metaphor processing requires theextension of the meaning of the metaphor to include the contextas Glucksberg and Keysar (1990) suggest, the process would havealready taken place for approximately half the number of timesthe target appeared in a LITERAL context (i.e., those times itappeared as a metaphor first). The literal uses would then besomewhat easier to process. The contexts required no expansionthemselves, since they provided an abundance of relevantinformation. The difference in the present experiment is thatMetaphor Comprehension - 119the contexts did not provide enough information to constrain thetarget selection, and so rather than constraining interpretationof the targets, what was required instead was an elaboration ofthe contexts. Such elaboration could only occur post hoc, thatis, after exposure to the target choices. Now the advantage isreversed, since the likelihood of the contexts already havingbeen encountered with a target of the same meaning was 3:1 (theSURPRISE condition, while sharing the same context, would notcontribute to the effect). Such an advantage would likely havebeen distributed over the three conditions (NONLITERAL,PARAPHRASE, UNFAMILIAR) to the point that it was undetectableindividually, but appeared as a relative handicap for LITERALresponses. Put another way, each time a LITERAL or NONLITERALtarget occurred, a different context would have to be elaborated,since the meaning of the target in that context was different,while the process would only need to occur once for the 3repeated conditions, since the meaning of the target in thatcontext was the same. Additional support for this possibility isprovided by the fact that the dead portion LITERAL resultsremained consistent with those of the earlier experiments, andalmost identical to results in the NONLITERAL condition (1138 and1144 msec, respectively). This result is exactly what would bepredicted by the argument that dead metaphors represent a case inwhich a class-inclusion had been established previously, throughrepeated exposure so the same broadened definition is availableto both usages. The high rate of errors in the SURPRISEMetaphor Comprehension - 120condition is also consistent with this explanation, since it alsorepresents a case where a unique meaning must be understood.The hypothesis that novel metaphors actually promoteunderstanding when context is reduced is not supported by theseresults of this experiment. The dead metaphors in the NONLITERALcondition were not notably different from the PARAPHRASE resultsin terms of response times (1144 msec and 1178 msec,respectively) although errors were somewhat higher. In contrast,the response times to novel metaphors in the NONLITERAL conditionwere now even slightly longer than SURPRISE condition responses(1588 vs. 1549 msec) and over 400 msec longer than the comparablePARAPHRASE condition (1182 msec). Further the error rate fornovel NONLITERAL responses was dramatically higher whencontrasted to the PARAPHRASE condition (51 vs. 4). In fact,12.75% of all responses to novel metaphors were wrong choices.Suggesting that metaphor provided additional informationfrom which to construct a complete meaning seems plausible untilthe error results of Experiment 6 are considered. What couldaccount for such a result? The first, and most obvious option issimply that Ortony et al. (1978) were right - that subjects needa long, well developed context to understand metaphors. In thiscase, it is worthwhile to point to the NONLITERAL results in thedead metaphor portion and emphasize that even if Ortony et al.are correct, such a finding apparently only applies to novelmetaphors, since the two portions behave in quite different ways.A second option is to consider these results as somehowMetaphor Comprehension - 121idiosyncratic to a few subjects. However, in Experiment 6 everysubject made at least one error in the NONLITERAL condition, and6 of 20 made 4 or more, while in Experiment 5, 14 of 20 subjectsmade no errors in this condition, and none made 4 or more. Thisappears to indicate a consistent effect across subjects, but suchresults would gain reliability if they were found to replicate ina new study, preferably with an entirely different stimulus set.It is probably true that at least some of the effects areattributable to the specific stimuli used. With theseconsiderations in mind, a third possibility may be raised.According to Glucksberg and Keysar (1990), metaphor arises when atopic is included in a class to which it had not previouslybelonged, because features of a hierarchically superior categoryget extended from the basic level term. In the example "My jobis a jail", the characteristics of a jail are extended to includemy job, which in 'literal' usage would not qualify for inclusionin such a category. When presented with a very underdevelopedcontext, it is likely that subjects are unable to make use of therelevant metaphorical class inclusion structure, particularlywhen they are unsure of which of the two targets requireexpansion in the first place. Consider as an example onePARAPHRASE item from the experiment as shown in Figure 12. Whenwe have read a context like "They complained his work had become"and the target choices are "sand" and "dull", we can make achoice even though context information is limited. The onlycontent word in the context is "work" and we are far better ableMetaphor Comprehension - 122to develop a class inclusion for "work" with "dull" than with"sand". The choice in the NONLITERAL condition for the samecontext is between "tapestry" and "stagnant". The correct choiceis now much less apparent. Work may have some elements in commonwith stagnation, but it might also be possible to find ways thatwork resembles tapestry. It has been said that we can find somecommonality between any two things, if we try hard enough. Thismay be what is happening here. The struggle and ultimateguesswork involved in resolving this choice is reflected inlonger response times and more errors.Now, compare the same stimulus context as it appears inExperiment 5. It reads: "Edmund had written two novels before heturned 30. Later he wrote three more, but his editors rejectedthem all. They complained that his work had become". Given thesame two choices, the preference for "stagnant" over "tapestry"is unequivocal. We realize that the work referred to is writingnovels, that rejection of later work is involved, and we can seethe consistency between this contextual implication and thefeatures of the word "stagnant" that refer to a once attractiveentity that has lost its vitality, to a dullness and lack ofprogressive movement, to a lifelessness and lack of appeal.Instead of forcing a connection from word to context, the choiceis between a word that appears patently nonsensical and one thatnow seems not only eminently suitable, but rich with appropriateresonances.It also seems pertinent to the discussion to think about theMetaphor Comprehension - 123target choices from the point of view of language use generally.The recent attention to the study of pragmatics has shown howmuch the derivation of meaning from language consists ofcontextual extrapolation. Developmental studies of languagelearning (e.g., Newport, Gleitman, and Gleitman, 1977) haverevealed that children must constantly constrain meaning inalmost miraculous ways in order to connect what is meant to whatis said. In such circumstances, it is essential to be open tomeaning, rather than searching for closure. At its extreme, weequate 'closed-mindedness' with 'literal' thinking and anunintelligent denseness. It is the impression we get whensomeone doesn't pick up on our irony, or the embarrassment wefeel when we realize we have been unable to detect someone else'ssarcasm toward us. The situation created when context isseverely limited means that there is no certainty of closure.The results of this experiment support Ortony et al.'s(1978) conclusions in one way, but not in another. Shortcontexts produce longer response times, and affect theunderstanding of novel (but not dead) metaphors more than literalwords. Understanding appears to require abstracting informationfrom the metaphor to fit the context as much as it requiresabstracting information from the context to fit the metaphor.Perhaps the most surprising finding from this experiment isthat response times in the majority of conditions were no longerthan those of Experiment 5. This result is elaborated on in thenext section, but it is interesting to discover that even a veryMetaphor Comprehension - 124limited context can provide enough information to permitunderstanding when class inclusion is unproblematic.Metaphor Comprehension - 125BETWEEN GROUPS ANALYSIS OF EXPERIMENTS 5 and 6In Experiment 6, information from the Experiment 5 contextswas reduced by changing nouns to pronouns and eliminatingextraneous, but schemata enhancing, details. The same orders ofpresentation were used in both experiments. This design resultedin the presentation of a long and short version of every stimulusthat permitted a between-groups analysis with experiment (i.e.,context length) as an additional factor. The comparisonresembled Ortony et al.'s (1978) metaphor experiment in whichdifferent subjects received either long or short contexts for anentire experiment. Further, by combining the two experiments(Experiment 5 and Experiment 6), the total number of subjects wasthe same as Ortony et al.'s (40) although many more stimuluspresentations per subject were made (200 as opposed to 16). Itwas expected that overall response times, time response errors,and target choice errors would be greater in Experiment 6 than inExperiment 5, given that contexts were reduced to approximatelyone-quarter their former length, and that context hasconsistently been recognized as an important component inlinguistic understanding. However, when the response timeresults of Experiment 5 and Experiment 6 were compared in a 2 X 2X 5 - experiment (5 vs. 6) by portion by condition -between/within ANOVA, there was no statistically significantdifference between the overall mean response times of the twoMetaphor Comprehension - 126experiments [F(1,38) = .02, p = .878]. Although there weresignificant reversals in some conditions (to be discussed), theoverall means for the two experiments differed by only 13 msec(1391 msec and 1404 msec, respectively). The number of deletionsbecause of responses of either less than 100 msec, greater than10,000 msec, or two standard deviations beyond the subject's meanfor that condition were almost identical (230 in Experiment 5 and235 in Experiment 6), but the number of response errors washigher in Experiment 6 (437 as opposed to 381). Reducingcontextual information may have specific effects onunderstanding, but it does not appear to have had any overalleffect on how long it took to choose a target and respond to itin the semantic choice task used here. Table 2 provides acomparison of response time violations and error rates for all ofthe conditions in the two experiments.The same ANOVA also showed that the experiment (5 vs. 6) byportion by condition interaction was only marginallystatistically significant [F(3.1,117.3) = 2.38, p = .070].Figure 13 makes this effect apparent. Of particular interest isthe LITERAL condition, which was much slower when context wasminimal, and the SURPRISE condition, which unlike thedisadvantage it suffered in Experiment 5, was faster than theUNFAMILIAR condition in Experiment 6, and even slightly fasterthan NONLITERAL novel metaphors. Reducing contextual informationappears to have had little effect on the processing of UNFAMILIARwords, and to have actually sfleedkresponding -imrthe SURPRISEMetaphor Comprehension - 127condition for both dead and novel portions. The same pattern ofresults as the five condition analysis was found when just theNONLITERAL, PARAPHRASE, and LITERAL conditions were assessed. Nostatistically significant difference was found between the twoexperiments [F(1,38) = 2.09, p = .156], and the experiment byportion by condition interaction was not significant [F(1.3,50.8)= 1.09, p = .320]. Results of Experiments 5 and 6 may becompared in Figure 13.Patterns of errors between the two experiments were alsoexamined. These are shown in Figure 14. The ANOVA for overallmean errors between Experiments 5 and 6 (long vs. short context)showed no significant difference [F(1,38) = 2.67, p = .11], butthe experiment by portion by condition interaction wassignificant [F(3.1,119.5) = 11.54, p < .001]. The error patternthat occurred in the novel portion of Experiment 5 may accountfor much of this difference.It is also interesting to contrast patterns of results formean response times with patterns for mean errors. Clearly, theinformation they provide is quite different. In particular, theerror rates for novel metaphors in the reduced context experimentshow a dramatic rise over all of the other NONLITERAL conditions,and the SURPRISE conditions for Experiment 6 shows more errorsthan the UNFAMILIAR conditions, even though the SURPRISEcondition response times are faster.One expectation of shortening the contexts is that subjectsmight make greater use of extra-contextual information, including■ NOVEL6DEAD6O NOVELS• DEADS200010001500Metaphor Comprehension - 128MEANS - EXPERIMENT 5 AND 6N P^L^S^UCONDITIONSFigure 13: Results of Experiments 5 and 6 together - Dead and novelportions in a semantic decision tasli. (N)nonliteral (P)paraphrase (141iterat(S)surprise (U)unfamiliar.■ NOVELSD DEAD6El NOVEL5■ DEAD5302Metaphor Comprehension - 129ERROR MEANS - EXPERIMENT 5 & 6N P^L^S^UCONDITIONSFigure 14: Error means for Experiments 5 and 6 together - Mean number oferrors per subject for dead and novel portions in a semantic decision task.(N)nonliteral (P)paraphrase (L)literal (S)surprise (U)unfamiliar.Metaphor Comprehension - 130what could be gleaned and retained from previous presentations ofthe same stimulus. This would be most clearly demonstrated in adifference between first and last presentations in repeatedcontexts having a consistent meaning. However, when the meansfor first vs. fourth presentation for the various conditions wereexamined, only the UNFAMILIAR condition showed a clear trend, andthis was no more pronounced in Experiment 6 than Experiment 5.Nevertheless, this effect was very large. First presentations ofUNFAMILIAR targets averaged nearly 300 cosec slower than fourthpresentations. The order ANOVA between Experiments 5 and 6together for the three base conditions (with the SURPRISE andUNFAMILIAR excluded) shows no significant interaction of orderwith the experiment by portion by condition results [F(1.6,59.4)= .07, p = .888]. First appearances of a condition were notresponded to significantly slower than fourth appearancesoverall. However, the even/odd analysis showed that arbitraryhalf did interact with the experiment by portion by conditionresults in the base conditions to a statistically significantextent [F(1.6,59.8) = 5.58, p = .01].UNFAMILIAR conditions may be particularly disadvantagedsince it is likely that neither target word can provide much helpin understanding the meaning of the context (recall that bothtarget and nontarget were always matched for word frequency).However, it might also be predicted that the developed context ofExperiment 5 would consequently be more useful in understandingthe UNFAMILIAR targets- The _results de-not - -support thisMetaphor Comprehension - 131expectation. Experiments 5 and 6 do not differ appreciably inoverall response time for their respective UNFAMILIAR conditions.It appears that if a word is unfamiliar, the information providedby the immediate context is not a great help in understanding it.The speculation that the LITERAL condition takes longer whencontext is short because contexts must be expanded post hoc tofit the targets was advanced in the discussion of the Experiment6 results. If we accept this proposal, a neat explanation forthe SURPRISE results emerges. The SURPRISE condition is the onlycondition similar to the LITERAL in the sense that the meaningdeveloped by the repeated context, and consistent over threeother conditions, has become inappropriate. The conditions arealso similar in that they consist of a context followed by aliteral word ending. The results section of Experiment 5discusses why the SURPRISE targets would require a post hocrevision of the contexts. Since the other conditions have longcontexts and do not require this procedure, the SURPRISEcondition is substantially handicapped. In Experiment 6, theSURPRISE condition still takes longer than the base conditions,and accounts for a substantial portion of errors. However,relative response times are faster in Experiment 6. Possibly,the 'double take' revision of contextual meaning that took placein Experiment 5 is of little use when contextual information isalready diminished. The closure that accounted for the surpriseeffect does not occur, and so revision of contextual expectationsis less importa___14nw-il—een-te)cttrar -b-5-c-p-etions are weaker, theMetaphor Comprehension - 132subject is more open to alternative possibilities. Another wayto say this is that the SURPRISE targets were simply lesssurprising, because shortening the contexts made subjects lessrigid in their expectations.The question of why subjects could respond to stimuli justas quickly in Experiment 6 as Experiment 5 is truly puzzling, andmerits further investigation. The only, highly tentative,explanation that seems plausible is that the cognitive approachcalled for by the lack of closure when context is reduced gives aresponse time advantage, independent of any advantage conveyed bysolid information. This hypothesis will be further developed inthe General Discussion to follow.Metaphor Comprehension - 133GENERAL DISCUSSIONThree notable findings relating to metaphor understandingemerge from the present series of experiments. The first is thatmetaphor comprehension appears to be so dependent uponlexicalization - not of a word per se, but of the specific waythe word is used - that it is impossible to draw conclusionsabout the distinction between literal and metaphorical fromresponse time studies without simultaneously considering itseffects. Variations in familiarity can reasonably account forall the response time differences found in these within-subjectsmetaphor experiments and can explain why, in various empiricalstudies, different forms of nonliteral language (idioms,metaphors) have shown differing relative response rates whencompared to literal words. This is not to say that metaphoricallanguage is in no way different from literal language, but onlythat it is probably not different in terms of demanding adiscrete processing stage where inappropriate literal meaningsare tested against context and subsequently rejected and revised.While these experiments refute the implication of stagemodels that every metaphor should show a longer response time,they have demonstrated that a temporally expensive process ofmatching meaning to context is taking place for unfamiliar words,or familiar words used in a surprising or unfamiliar way, withnovel metaphors_being---a—eliaracter-i-g-t-i-a —example of the latter.Metaphor Comprehension - 134On the other end of the familiarity continuum, deadmetaphors can be as familiar as so-called literal words, and cantypically be accessed just as quickly. The lexicalized metaphoris still a metaphor and as such represents a prototypicalexemplar of a class. The difference is that the development andexpansion of the ad hoc class it represents is not required sincedead metaphors, like literal words, have categorical definitionsthat are already lexicalized. In terms of language, this meansthat multiple uses are known and available to the user. Both maybe considered exemplars of different aspects of the same meaning,and they would likely be represented in a dictionary as primaryand secondary meanings of a single word.The second result of note involves the contribution ofcontextual preparedness to comprehension. Context is a verysignificant factor in shaping expectations for understanding whatis to follow. In the present study, targets with meaningssupported by context (in the PARAPHRASE condition) wereconsistently processed more quickly than targets with meaningsunsupported by context (in the SURPRISE condition). Sincetargets in both these conditions were literal words matched forfamiliarity in an identical context, the differences inprocessing time must have to do with the fact that the meaning ofthe SURPRISE targets is less likely to be anticipated.The ability of context to constrain possible alternativemeanings is both a strength and a weakness. As a strategy it isparsimonious tot_e_abla----te—anticipate toih—at as to come. Such anMetaphor Comprehension - 135approach enhances both speed and depth of understanding. At thesame time, we could virtually never absorb anything new if wewere always able to anticipate what we are about to hear.Context would be detrimental if it constrained language to such adegree. Experiment 4 shows the error of expecting that meaningcan be fully constrained by context, even when that contextprovides a fair amount of information. The final two experimentsreveal that overall comprehension speed is not affected by a lessdeveloped context, but that certainty of understanding may beweaker. Perhaps metaphor intervenes in the comprehension processby forcing us to remain flexible. We are particularly inclinedto reserve judgment when we encounter a novel metaphor in asituation where context has not provided sufficient constraintsince we might be wrong - but we know we might be wrong, and thisstrategy allows us to make a graceful recovery as context becomeselaborated.To illustrate, consider the metaphor "Juliet is the sun"first cited by Searle (1979) and also employed by Chandler(1991). This phrase is impossible to understand in isolation andwithout reflection. If we have minimal context (say, what we canremember about Shakespeare's Romeo and Juliet - and assuming thatwe have even had such exposure) we might be able to make somesense of it if we can recognize it was spoken by Romeo, who wasinfatuated with Juliet. Reading the play and encountering theline as Romeo alludes to the sun at daybreak, when the moonbecomes pale by__colaparlson-r--we -can-laelin to see ways that the sunMetaphor Comprehension - 136has features that Romeo could apply to Juliet. When we fullyexamine the play, we find that it is permeated with light andcosmology imagery, providing a considerable body of context.However, it is arguably really only Romeo (or Shakespeare) whocan truly understand what he means when he utters the metaphorbecause it is only he who can know what he has encoded in theclass of sensations he associates with the sun. If we are tomake progress toward ecological validity in metaphor research, wemust recognize that for a subject in a psychology experiment toparaphrase what a metaphor means, and for a psychologist toevaluate such a description, moves well away from directlyaccessing the cognitive processes of language understanding.The third area into which the preceding experiments havebeen able to provide insight concerns the question of how thetarget words worked together with context to create meaning.Experiments 1, 2, 3, 5, and 6 each had stimuli in whichconditions had identical contexts but different endings. Amongthese, novel metaphors took less time to understand thansurprising or unfamiliar words, and dead metaphors were virtuallyno different than literal words once word frequency wascontrolled. A particular strength of the class-inclusion modelis that it provides a reason why metaphor produces thisadvantage. When we are processing new information, metaphorpermits that information to be expressed in a way that we canaccess through what we already know. Through the mutuallycompatible processes of -earrstrairit from context and enhancementMetaphor Comprehension - 137from metaphor, the likelihood of successful communication isgreatly increased.The particular suitability of using metaphor for providingnew information raises an important point about feature-matchingor 'topic and vehicle' theories in general. As Glucksberg andKeysar (1990) point out, theories that depend upon matchingfeatures of the topic with those of the vehicle, whether byinteraction (Black, 1979), salience imbalance (Ortony, 1979),similarity (Tversky, 1977), common ground (Verbrugge andMcCarrell, 1977) or interactive competition (Chandler, 1991)assume that the features of the topic are as well known as thoseof the vehicle. If this were true, metaphor would only beredundant, not informative. Saying "Mary is a kitten" if we knowthat Mary is charming and playful only confirms what is alreadyknown. Saying "Mary is a kitten" if we don't know anything atall could very well mean that we are discussing a household pet.It is only if there is an intent to inform - if we know Mary is ahuman and we are discussing her appeal - that the metaphorbecomes useful. This imbalance between metaphor as the source ofinformation and topic as the recipient of it is important, andhas not been formally developed. Metaphors provide or confirmfeatures about the topic, overwriting default values forunderdeveloped schemas with specific information. For thisreason, all successful metaphors implicitly depend upon some sortof pre-established contextual framework.Returning_totne_T.Juliet-is e^-sunit-example, featureMetaphor Comprehension - 138matching theories would find difficulty with this metaphor, sincethere is very little interaction on a feature-by-feature analysisthat could yield a foundation on which to provide information.When Glucksberg and Keysar (1990) claimed that metaphoricalcomparisons are semantically asymmetrical and fundamentallyirreversible, the explanation would seem to be because thefeatures of the metaphor relevant to the information beingconveyed are always better known than those of the topic. Thisis why we must know a literal meaning of a word in order for usto understand the metaphor derived from it. Such a conclusionsupports the contention that literal meaning is fundamentallyrelated to familiarity with a specific usage and represents whatI believe literal meaning to be.Glucksberg and Keysar's (1990) attempt to draw a distinctionbetween what is literal and what is metaphorical runs into aproblem which I believe the class-inclusion model cansuccessfully explain. They claim that one clear differencebetween the two is the availability of the class-inclusionconstruction for metaphoric, but not literal, comparisons.Metaphoric statements, which compare two unlike things, can beparaphrased to look like class-inclusion statements (My job islike a jail/ My job is a jail), while literal statements, whichcompare two like things, can't be paraphrased in this way (Beesare like hornets/ *Bees are hornets). The problem with thiscontention is that there would appear to be just as many literalstatements that_am_r—lass—ineittsloTi --ttder, but they can't beMetaphor Comprehension - 139paraphrased as similes (Bees are insects/ *Bees are likeinsects). The explanation for this effect does not lie in howalike the things are, but rather in the fact that all class-inclusion statements involve categories at different levels,while all similes involve categories at the same level. Becausethe class-inclusion model uses the same metaphorical word as bothtype and token, the metaphor can be referred to on either level.That is, if we say "My job is like a jail", we are using "jail"as a category at the same level as "my job". If we say "My jobis a jail" we are using "jail" as a prototypical exemplar (and defacto name) of a superordinate category. It might still bereasonable to state that only metaphor has this linguisticflexibility and that in contrast, literal statements may only beparallel or hierarchical, but never both.It is easiest to think of a metaphor as providing newinformation, but we can easily think of examples where this isnot the case. If we both know Mary equally well, and we bothknow that she is charming and playful, then it could be arguedthat saying "Mary is a kitten" is not providing new information.However, if such a statement were uttered, its intent wouldlikely be to elicit confirmation of the speaker's observationwith a second opinion. This is new information of a somewhatdifferent kind, but it also illustrates a useful linguisticfunction fulfilling a common human need.Research on idiom and proverb understanding (Hoffman and_Kemper, 19871 has_shown-that priming with one proverb helps inMetaphor Comprehension - 140the recall of another with the same meaning, and that an initialfocus on the image of the proverb can detract from comprehensionof the implicit figurative meaning, while an initial focus on themeaning makes the images subsequently more meaningful. Thisabstraction of meaning is consistent with Bartlett's (1932)findings about subjects retaining the gist of stories, whileadapting the details to previously formed expectations. Thisseems to be essentially what is happening on a different scale inthe case of semantic priming, when more abstract categoricalsimilarities are recognized. If the same process happens with aword when it is used metaphorically, the gist of the word in itsliteral sense provides a bootstrap in the form of a categoricalinclusion into an abstract, thematic class which can makecomprehension more complete. No feature matching is necessary.Only the 'literal' meaning of the metaphor needs to be jointlyunderstood. The similarity between Glucksberg's (1991)hierarchically superior metaphorical class, Shank's (1982)abstract, thematically structured TOPs, and Rosch's (1973; 1978)superordinate level categories is readily apparent. If a TOP-like structure can explain semantic linkages between stories andexperiences, it can also explain why concepts like 'doctor' and'hospital' can be semantically linked, and why we can employ asimilar process in understanding what is meant by "My job is ajail".Some recent challenges to Glucksberg and Keysar's (1990)class-inclusion model lave---come froffi -Gibbs (1992). Gibbs'sMetaphor Comprehension - 141criticisms focus on the belief that metaphor is a fundamentalcharacteristic of how people categorize and make sense of theirexperience. Drawing on Lakoff's (1987) ideas concerning imageschemas, Gibbs asserts that the process of understanding metaphorusing ad hoc superordinate categories is rare. More usually, hecontends, metaphors derive from concepts already existing inlong-term memory. He cites "Love is a journey" as an example ofa fundamental structural metaphor.While Gibbs's assertion that metaphor usage reflects afundamental process related to categorization and meaning may becorrect, Glucksberg, Keysar and McGlone (1992) have producedempirical evidence that subjects interpreting metaphors are notconstrained by basic image schemas. It seems unlikely thatsomething like the "love is a journey" metaphor is inherentlyfundamental, for many of the same reasons why any proposal ofinnate concepts seems problematic. We can certainly understandthe meaning of a metaphor couched in an infinite number ofcompletely independent conceptual frameworks; i.e., "Love is arose" or "Love is a warm puppy". Virtually all investigations ofmetaphor in some way allude to its very ability to restructurefamiliar things in creative ways - essentially a definition ofthe process of ad hoc categorization. Systematicity in metaphor,in my view, arises more from the conventionalization process (aparticularly apt metaphor becomes culturally widespread and thengenerates variants on a consistent theme) than from innate oreven cultural constraints- Thc-falyt-that th-éCiass-inclusionMetaphor Comprehension - 142model specifies that the conceptual class framework accessed by ametaphor is previously unnamed provides a model for how theinteraction between thought and language is achieved. Lakoff'smodel would appear to depend upon foundational metaphors beingrigidly lexicalized, an idea that the preceding experimentsappear to refute.Although the model presented by Glucksberg and Keysar (1990,Glucksberg, 1991) provides what seems to me the most viableexplanation of metaphor understanding to date, and one that issupported by the data of the experiments presented here, thereare some points of their theory that in my opinion could bestrengthened by elaborations suggested by the present studies.Glucksberg and Keysar (1990, Glucksberg, 1991) decisively rejecta three stage model like Clark and Lucy's (1975) (i.e., theStandard Pragmatic Model), citing the evidence of Ortony et al.(1978), Gibbs (1984), and others who have shown that metaphors incontext need take no longer to understand than comparable literalsentences. The reader is left with the implication that a class-inclusion model would predict that metaphors and their literalcounterparts are processed equally quickly. However, in Ortonyet al.'s experiments, the finding of 'no difference' occurred foridioms in one experiment, and metaphorical phrases in elaboratedcontexts in another. In most of Gibbs's experiments, idioms orindirect requests were studied and then generalized to explainall nonliteral language. Both bodies of work focus on uses ofnonliteral language-that-----from the point of view of the presentMetaphor Comprehension - 143studies appear to be special cases. While the present resultssupport the class-inclusion model as an explanation of theprocess involved in understanding metaphor, it seems consistentto suggest that when the process of expanding a superordinateclass in order to encompass a new member is required, such aprocess takes more time than when it is unnecessary - that is,when the class into which the inclusion takes place is alreadynamed and lexicalized.In advancing this explanation I am not suggesting that adifferent sort of three-stage model is called for. We do notalways need to test and expand a hierarchical class when anymetaphor is encountered, as the dead metaphor resultsconsistently show. What I am saying is that there is a continuumof latency depending on whether or not we need to 'enlarge ourunderstanding' as it were, and the enlarging process takes time.The time required appears to be related to how familiar we arewith the meaning of the word used metaphorically.The second qualification to the work of Glucksberg andKeysar (1990) comes from the importance of context tounderstanding, and to understanding metaphors in particular.Context is not simply a matter of a single topic word for ametaphorical vehicle. The hypothetical "a is b" metaphorvirtually never exists in isolation. It is important torecognize that context is always working to constrainpossibilities, so much so that a very well developed context willlikely result in_am interpretatiorviar metaphor by default, asMetaphor Comprehension - 144Ortony et al. (1978) suggest.A tentative model for a theory of metaphor understandingbased on a combination of context and word or phrase target canbe described by a lock and key analogy, with context representedby the lock and the various target possibilities by the keys. Toillustrate how this model would work, consider the samplestimulus item from the novel metaphor portion of Experiment 5presented in Figure 12. In the context of a promising youngwriter whose works are now being rejected, the PARAPHRASE target"dull" is well supported by the context, although any of a numberof other choices with different meanings could fit as well, suchas repetitive, boring, or dated. In the analogy, the context hasprovided a clearly defined but undifferentiated keyhole that isso generic that even if the intended key were missing, we couldfind another one that would fit adequately. However, if thecontext were to become very highly developed, the keyhole wouldbecome reciprocally more constrained, finally reaching the pointwhere only a single key would be likely to fit. When we have acontext that shapes the target to this extent, we can recognizethe appropriate key even in its absence, simply by the shape ofthe keyhole. This anticipatory advantage makes the searchprocess highly efficient.In the case of the UNFAMILIAR target, "hackneyed" is a wordwith a similar meaning that is usually understood but less oftenused. It fits the context, but takes longer to process becauseit is less familiar— We-knew-whwt - itloOks like, but we have toMetaphor Comprehension - 145search the key ring before we can locate it, and some of us mightsearch for quite a while but be unable to find it. The SURPRISEtarget "childish" also takes longer than the PARAPHRASE, but fora different reason. It has a well lexicalized meaning, but thatmeaning does not immediately appear to fit the context. TheSURPRISE meaning is a strong key that reshapes the contextkeyhole to make itself fit. The same schemata are present, butdifferent parts are involved. We focus less on the fact that theauthor was successful in the past, and more on his success beingtied to his youth. Reshaping the keyhole also takes time. Thelexical meaning of the NONLITERAL target "stagnant" seems not tofit the keyhole at all, but one aspect of it fits the sameportion of the context as does "dull". Abstractly, both "dull"and "stagnant" can be members of a class that might becharacterized as static or uninteresting. In addition, however, arich web of other associations in the class typified by stagnant(but not by dull) like stale, unmoving, and overripe, along withnonverbal sensory ones, enrich our understanding of how theeditors regarded the author's work. The metaphorical key notonly fills the keyhole completely, it overfills it, expanding thecontext to take it all in.Some of the questions raised by the Gibbs (1984, 1989) -Dascal (1987 - 1989) debate over whether literal meanings existat all are clarified by the results obtained in these studies.Briefly, Gibbs claimed that literal meanings do not havepsychological reality -s4nee-meaning is not simply a function ofMetaphor Comprehension - 146compositional analysis and since his results have shown that aliteral analysis stage is not obligatory in the determination ofmeaning. Dascal argued that literal meaning can be equivalent toconventionally established meaning, and in this sense, literalmeaning both does exist and does participate in the determinationof speaker meaning. By equating what is ordinarily defined asliteral meaning with conventionality, both author's points areaccommodated. Within the framework of literal and metaphoricalthat he employs, Gibbs is correct in saying that literal languageis not always accessed before a nonliteral interpretation ismade, as his own results with idioms show. On the other hand,consideration of novel metaphors confirms Dascal's pointregarding the need to determine meaning when a usage is notconventionalized. Then the literal meaning of the word must playa role. The longer response times to novel metaphors in thepresent experiments demonstrate the effects of this process.Perhaps the simplest way to sum up the findings of thestudies developed for this dissertation is to say that theypresent metaphor as much less mysterious than many conceptionsconvey. If we see metaphors as words or phrases that meansomething completely different than we want to say, they seem abizarre, irrational aberration of language. If instead we seemetaphors as words or phrases that convey new information in away that makes it more easily comprehensible, they seem aflexible, efficient tool of language. This approach to metaphoris remarkably consisterit-with-the---AzMy- we understand otherMetaphor Comprehension - 147cognitive processes, like learning, memory, and concept formationto operate. While the actual processes involved incategorization and generalization that permit this to happenremain speculative, the results of these studies unequivocallypoint toward the functional view.Metaphor Comprehension - 148CONCLUSIONThe present experiments have shown that three factors:familiarity of usage, contextual preparedness, and conceptualclass, represented by the literal meaning of words usedmetaphorically, all contribute to understanding any discourse.Familiarity effects can account for the processing delaysassociated with nonliteral language. Since identifyingnonliteralness with unfamiliarity of usage can explainmetaphorical effects, the question of what possible functionmetaphor might have then occurs. In reply, I have proposed thatwhen context is adequate but the concept is new, complex, unique,or superlative, using metaphor provides a way of approaching theidea through class-inclusion. By employing what is known about afamiliar class to provide informational depth to a new one,metaphor is given a functional explanation. When context isinadequate, metaphors are much less useful but possibly serve toalert us to seek extracontextual cues. The time cost ofunderstanding metaphors is compensated for by the quality ofinformation they can convey. If metaphors provide us access toinformation that would otherwise be unknown and inexpressible,the additional time it takes to understand them seems a smallprice to pay.Metaphor Comprehension - 149REFERENCESALLMAN, W.F. (1989). Apprentices of Wonder: Inside the Neural Network Revolution. New York: Bantam Books.ALLEN, P.A., MCNEAL, M, & KVAK, D. (1992). Perhaps the lexiconis coded as a function of word frequency. Journal of Memory and Language, 31, 826-844.BARTLETT, F.C. (1932). Remembering: A Study in Social Psychology. London: Cambridge University Press.BLACK, M. (1979). More about metaphor. In A. Ortony (Ed.)Metaphor and Thought. New York: Cambridge University Press.BLANK, G.D. (1988). Metaphors in the lexicon. Metaphor andSymbolic Activity, 3(1), 21-36.BOBROW, S. & BELL, S. (1973). On catching on to idiomaticexpressions. Memory and Cognition, 1, 343-346.CHANDLER, S.R. (1991). Metaphor comprehension: A connectionistapproach to implications for the mental lexicon. Metaphor and Symbolic Activity, 6(4), 227-258.CHOMSKY, N. (1965). Aspects of the Theory of Syntax.  Cambridge:M.I.T. Press.CLARK, H.H. & LUCY, P. (1975). Understanding what is meant fromwhat is said: A study in conversationally conveyed requests.Journal of Verbal Learning and Verbal Behaviour,  14, 56-72.DAMASIO, A.R. & DAMASIO, H. (1992) Brain and language. ScientificAmerican, 267(3), 88-95.DASCAL M. (1987). Defending Literal Meaning. Cognitive Science,_11, 259-241.Metaphor Comprehension - 150DASCAL, M. (1989). On the Roles of Context and Literal Meaningin Understanding. Cognitive Science, 13, 253-257.FODOR, J.A. (1975). The Language of Thought. Cambridge, MA:Harvard University Press.FODOR, J.A. (1983). The Modularity of Mind. Cambridge, MA: MITPress.FORSTER, K.I. (1990). Lexical processing. In D.N. Osherson & H.Lasnik (Eds.) Language: An invitation to cognitive science,Vol. 1. Cambridge, MA: The MIT Press.FRANCIS, W.N. & KUCERA, H. (1982). Frequency Analyses of English Usage: Lexicon and Grammar. Boston: HoughtonMifflin.GARROD, S. & SANFORD, T. (1988). Thematic subjecthood andcognitive constraints on discourse structure. Journal of Pragmatics, 12, 519-534.GIBBS, R.W. (1979). Contextual effects in understandingindirect requests. Discourse Processes, 2, 1-10.GIBBS, R.W. (1980). Spilling the beans on understanding andmemory for idioms in conversation. Memory and Cognition, 8,449-456.GIBBS, R.W. (1984). Literal meaning and psychological theory.Cognitive Science. 8, 275-304.GIBBS, R.W. (1989). Understanding and literal meaning.Cognitive Science. 13, 243-252.GIBBS, R.W. (1992). Categorization and metaphor understanding.Psychological Pev ew-^X52=-511.Metaphor Comprehension - 151GLUCKSBERG, S. (1991). Beyond literal meanings: The psychology ofallusion. Psychological Science, 2(3), 146-157.GLUCKSBERG, S. & KEYSAR, B. (1990). Understanding metaphoricalcomparisons: Beyond similarity. Psychological Review, 97(1),3-18.GLUCKSBERG, S., KEYSAR, B., & McGLONE, M.S. (1992). Metaphorunderstanding and accessing conceptual schema: Reply toGibbs (1992). Psychological Review, 99(3), 578-581.GRICE, H.P. (1975). Logic and conversation. In P. Cole & J.L.Morgan (Eds.) Syntax and Semantics: Speech acts. Vol 3.New York: Academic Press.HEIDER [ROSCH], E. (1972). Universals in color naming and memory.Journal of Experimental Psychology, 93, 10-20.HOFFMAN, R.R. & KEMPER, S. (1987). What could reaction timestudies be telling us about metaphor comprehension?Metaphor and Symbolic Activity, 2(3), 149-186.HUBBELL, J.A. (1988). The effects of metaphorical and literalprocessing on subsequent lexical latencies of sentencecomponents. Ames, IA: Unpublished doctoral dissertation.JASTRZEMBSKI, J.E. (1981). Multiple meanings, number of relatedmeanings, frequency of occurrence, and the lexicon.Cognitive Psychology, 13, 278-305.KINTSCH, W. & MROSS, E.F. (1985). Context effects in wordinterpretation. Journal of Memory and Language, 24, 336-349.KIRK, R.E. (1982). Experimental Design: Procedures for theBehavioral Sciences. Monterey, CA: Brooks/Cole.Metaphor Comprehension - 152KITTAY, E.F. (1987). Metaphor: Its Cognitive Force and LinguisticStructure. Oxford: Clarendon Press.KUCERA, H. & FRANCIS, W.N. (1967). Compositional Analysis of Present Day American English. Providence, RI: BrownUniversity Press.LAKOFF, G. (1987). Women, Fire, and Dangerous Things: What Categories Reveal About the Mind. Chicago: University ofChicago Press.LAKOFF, G. & JOHNSON, M. (1980). Metaphors we Live By. Chicago:The University of Chicago Press.Longman Synonym Dictionary. (1986). Essex: Longman Group, U.K.McCLELLAND, J.L. & KAWAMOTO, A.H. (1986). Mechanisms ofsentence processing: Assigning roles to constituents ofsentences. In J.L. McClelland, D.E. Rumelhart and the PDPResearch Group (Eds.) Parallel Distributed Processing: Explorations in the Microstructure of Cognition.  Vol 2.Cambridge MA: The MIT Press.NEUFELDT, V. (Ed.). (1988). Webster's New World Dictionary of American English. Third College Edition. New York:Webster's New World.NEWPORT, E., GLEITMAN, H., & GLEITMAN, E. (1977). Mother, I'drather do it myself: Some effects and noneffects of maternalspeech style. In C.E. Snow & C.A. Ferguson, (Eds.) Talking to Children: Language Input and Acquisition.  New York:Cambridge University Press.Metaphor Comprehension - 153NIMS, M.F. (Trans.). (1967). Poetria Nova of Geoffrey of Vinsauf.Toronto: Pontifical Institute of Mediaeval Studies.ORTONY, A. (1979). Metaphor: A multidimensional problem. In A.Ortony (Ed.) Metaphor and Thought. Cambridge: CambridgeUniversity Press.ORTONY, A., SCHALLERT, D.L., REYNOLDS, R.E., & ANTOS, S.J.(1978). Interpreting metaphors and idioms: Some effects ofcontext on comprehension. Journal of Verbal Learning and Verbal Behavior, 17, 465-477.POLLIO, H.R., FABRIZI, M.S., SILLS, A., & SMITH, M.K. (1984).Need metaphoric comprehension take longer than literalcomprehension? Psycholinguistic Research, 13, 195-214.POSNER, M. (1978). Chronometric Explorations of Mind. HillsdaleNJ: Erlbaum.RICHARDS, I.A. (1936). Metaphor. In I.A. Richards The Philosophyof Rhetoric. London: Oxford University Press.ROSCH, E. (1973). On the internal structure of perceptual andsemantic categories in cognitive development. In T.E. Moore(Ed.) Cognitive Development and the Acquisition of Language.New York: Academic Press.ROSCH, E. (1978). Principles of categorization. In E. Rosch &B.B. Lloyd, (Eds.) Cognition and Categorization. Hillsdale,NJ: Erlbaum.RUBENSTEIN, H., GARFIELD, L., & MILLIKAN, J.A. (1970).Homographic entries in the internal lexicon. Journal of Verbal Learning- and- Verbal Behavior. S, 487-49-4.Metaphor Comprehension - 154RUMELHART, D.E. & ORTONY, A. (1977). The Representation ofKnowledge in Memory. In R.C. Anderson, R.J. Spiro, & W.E.Montague (Eds.) Schooling and the Acquisition of Knowledge.Hillsdale, N.J.: Lawrence Erlbaum Associates.SCHANK, R. (1982). Dynamic Memory: A Theory of Learning in Computers and People. New York: Cambridge UniversityPress.SCHANK, R. (1990). Tell me a Story: A New Look at Real and Artificial Memory. New York: Charles Scribner and Sons.SEARLE, J.R. (1979). Metaphor. In A. Ortony (Ed.) Metaphor and Thought. Cambridge: Cambridge University Press.SWINNEY, D.A. (1979). Lexical access during sentencecomprehension: (Re)consideration of context effects.Journal of Verbal Learning and Verbal Behavior, 18, 645-659.SWINNEY, D.A. & CUTLER, A. (1979). The accessing and processingof idiomatic expressions. Journal of Verbal Learning and Verbal Behavior, 18, 523-534.TABOSSI, P. (1988). Accessing lexical ambiguity in differenttypes of sentential contexts. Journal of Memory and Language, 27, 324-340.TARTTER, V.C. (1986). Language Processes. New York: Holt,Rinehart, & Winston.TVERSKY, A. (1977). Features of Similarity. Psychological Review,84, 327-52.Metaphor Comprehension - 155VERBRUGGE, R.R., & McCARRELL, N.S. (1977). Metaphoriccomprehension: Studies in reminding and resembling.Cognitive Psychology, 9, 494-533.WHORF, B.L. (1956). An American Indian model of the universe. InJ.B. Carroll (Ed.) Language, Thought, and Reality: Selected Writings of Benjamin Lee Whorf. New York: Wiley.WITTGENSTEIN, L. (1953). Philosophical Investigations. (Trans.)G.E.M. Anscombe. Oxford: Blackwell.Metaphor Comprehension - 156APPENDICESMetaphor Comprehension - 157APPENDIX A158SENTENCE COMPREHENSIONOdie Geiger - Psychfest, 1990March, 1990This experiment involves the reading and understanding ofshort groups of sentences. On each trial, a few sentences willbe displayed, followed by a word or phrase at the end to completethe meaning. Your task will be to read the sentences carefully,and when you are sure that you understand their meaning up tothat point, push the spacebar at the bottom of the keyboard.Immediately, the sentences will disappear and the ending will bedisplayed. This portion of the experiment will be timed. Assoon as you understand the ending so that the whole presentationmakes sense, push the spacebar again. A new group of sentenceswill then appear on the screen and the procedure will continuethrough 160 trials.It is important to the success of the experiment that youunderstand the meaning of the entire presentation. In order toensure that you are reading the sentences carefully enough foryour responses to be meaningful, a short quiz will follow theexperiment, to find out if the meanings were clear to you. Thisdoes not mean that the sentences will be tricky or difficult.You will likely understand the meaning of the sentences as soonas you read them. Nor does it mean that you need to memorize thesentences, the details, or the names of people who may beincidentally mentioned. It is just a way of keeping you fromresponding automatically as soon as the final phrase is flashedon the screen.Since the initial portion of each presentation is not timed,you can leave it on the screen for as long as you like to makesure that the meaning is clear to you. After you press thespacebar, you should respond as quickly as possible, once themeaning of the whole is clear. We will provide a few practicetrials before the experiment begins to accustom you to theresponse procedure.You will have eight sets of 20 trials each to complete. Ifyou become tired at any point, please feel free to request abreak between sets. You will be given a point of class creditand an explanation of what the experiment is about after yoursessions have been completed.Thank you for your interest and participation.Odie GeigerJuliet ArmstrongResearch AssistantsIDIOMS1.APX^ 159EXPERIMENT 1 - IDIOMS1)II context: Dean spoiled the surprise that Joan had been planning fortheir mother's birthday. When he realized what he had done, heapologized for having1)II target: let the cat out of the bag.1)IP target: revealed the secret.1)IS target: fallen asleep in the closet.1)IL context: Walking back from the store, Anne found a kitten which sheput in with her groceries. She got home and her puppy went wildwhen she1)IL target: let the cat out of the bag.2)11 context: Watching the young men play football in the park on anearly spring evening made Ralph nostalgic. He found himselfwishing that he could2)11 target: turn back the clock.2)IP target: return to his youth.2)IS target: spend spring somewhere else.2)IL context: When he arrived an hour late for work Monday morning,Barry remembered that Daylight Savings Time had begun the nightbefore. As soon as he got home, he went into his bedroom to2)IL target: turn back the clock.3)11 context: Mary found it difficult to make friends at her new school.Finally, she decided to hold a party as a way to3)11 target: break the ice.3)IP target: become acquainted.3)IS target: try some new recipes.3)IL context: After running his team of horses for six hours over roughfrozen roads, Pierre arrived at an old barn with a watering troughoutside. His next task was to find a stick heavy enough to3)IL target: break the ice.4)11 context:George and Martha had a good relationship, because theyrecognized that neither was perfect. They both knew that Martha hada tendency to nag, and that George would occasionally4)11 target: fly off the handle.4)IP target: lose his temper.4)IS target: read yachting magazines.4)IL context: Benny was a tame parrot who was often allowed to fly aroundthe house. One of his favourite perches was right above the kitchensink. In order to turn on the water, it was first necessary to makeBenny4)IL target: fly off the handle.IDIOMS1.APX^ 1605)11 context: Curtis was an irresponsible gambler but he was saved by hicharming manner, and an incredible amount of good luck. No matterhow much difficulty he got himself into, he always seemed able to5)11 target: land on his feet.5)IP target: recover from his troubles.5)IP target: write scholarly essays.5)IL context: Mark had studied gymnastics since he was six. He was themost graceful athlete at his club, and the most talented. Hisroutines were both spectacular and controlled, and the audiencesknew that he would always5)IL target: land on his feet.6)11 context: When Mr.Wilson retired, he put his money into savings bondsAn investment counselor had suggested a way to earn more interest,but Mr. Wilson thought that he6)11 target: could smell a rat.6)IP target: might be cheated.6)IS target: should help the government.6)IL context: Linda hated the big old house she inherited. Every time sh4walked down the dirty, dingy hallway she heard noises and she wassure that she6)IL target: could smell a rat.7)11 context: The contact lens popped out somewhere on the basketballcourt. The referee had to stop the game while the players went overthe court7)11 target: with a fine-toothed comb.7)IP target: carefully and completely.7)IS target: with a vacuum cleaner.7)IL context: After swimming, Margaret's long hair was a tangled mess.She always had to spend extra time drying it, brushing it, andfinally, combing it7)IL target: with a fine-t000thed comb.8)11 context: The wedding was scheduled to take place in the garden thatafternoon. By 10 a.m., dark clouds began gathering. All the familycould do for the next four hours was to8)11 target: hold their breath.8)IP target: wait helplessly.8)IS target: finish remodelling the garage.8)IL context: One of the techniques Carmen had learned as a lab technicianwas that patients would become less nervous during an injection ifthey were told to look away and8)IL target: hold their breath.IDIOMS1.APX^ 1619)11 context: When her boss walked in her door with a big smile on hisface, Leslie knew she had been promoted. Weeks earlier, rumoursabout management changes had been9)11 target: in the wind.9)IP target: topics of speculation.9)IS target: about planned layoffs.9)IL context: The sky grew grey and the temperature dropped sharply. Hucraindrops splashed against the pavement, and piles of leaves swirlecabout9)IL target: in the wind.10)11 context: When John finally decided to do something about his lack cphysical fitness, he went to visit his doctor, who recommended aprogram that was10)II target:tailor-made for him.10)1P target:exactly appropriate for him.10)IS targt:designed for pregnant women.10)IL context: Oscar was tall and extremely thin, but despite this healways looked elegant and powerful in his suits because they were10)IL target:tailor-made for him.11)II context: All of Ben's earnings went toward financing his annualholiday in Reno, where he consistently11)II target: lost his shirt.11)IP target: gambled and lost.11)IS target: gave money to losers.11)IL context: The last time Ricky went swimming in the pond, a wind cameup and scattered his clothes. He finally found his pants, but hehad to confess to his mother that he had11)IL target: lost his shirt.12)11 context: Bob's parents and teachers had never had any trouble withhim as a child, but when he became a teenager he suddenly became12)11 target: too big for his boots.12)IP target: arrogant and disrespectful.12)IS target: expert at needlepoint.12)IL context: Carl's parents were shocked at how much it was costing tofeed and clothe their growing son. They bought him a new outfitfor school, but within only three months, he had become12)IL target: too big for his boots.13)11 context: The ad offered a 1968 Corvette in excellent condition for$2000. When the owner agreed to an offer over the phone of $1500,Cynthia hurried over to his house to13)11 target: nail it down.13)IP target: confirm the deal.13)IS target: see what color it was.13)IL context: The tornado caused considerable damage, but the onlydamage to the Johnson's house was a single loose shingle. -KrJohnson had to clinib up to the roof to13)IL target: nail it down.IDIOMS1.APX^ 16214)11 context: Charles disliked working with Neil, who wouldn't do hisshare and yet attempted to take credit for the completed project.Finally, Charles decided that it was14)11 target: a game two could play.14)IP target: a strategy he would try.14)IS target: wrong to dislike Neil.14)IL context: Jean and Sheila lived on a farm far away from childrentheir own age. When they received a new video game at Christmas,they were excited to find that it included14)IL target: a game two could play.15)11 context: Everybody had a good time at the Spring Dance; the factthat the caterers got lost and the food didn't arrive until 11 p.m.was the only15)11 target: fly in the ointment.15)IP target: disruptive event.15)IS target: reason Mary cried.15)IL context: Paula's opinion about the dubious joys of camping wasconfirmed after she got severely burned one day. When she went tofind something to relieve her agony, she discovered a15)IL target: fly in the ointment.16)11 context: The plant had suffered a number of financial losses asworld markets had changed. The employees didn't want to startlooking for other jobs, but they could16)11 target: see the writing on the wall.16)IP target: predict the consequences.16)IS target: organize a co-op.16)IL context: The kids had a great time scribbling while theirbabysitter was talking on the telephone. When they finished theywent outside, hoping that she wouldn't16)IL target: see the writing on the wall.17)11 context: All of Margaret's friends warned her against marryingBrian, but none were sympathetic when the marriage failed. They allfelt that Margaret had entered the marriage17)11 target: with her eyes wide open.17)IP target: knowing the likely consequences.17)IS target: for money and revenge.17)IL context: Tammy's friends wanted to see her even before she had fullyrecovered from the anesthetic, but the doctor suggested that shewould be more comfortable if she could visit them17)IL target: with her eyes wide open.18)11 context: Max had borrowed money to buy a house, a car, and a boatbefore he lost his job. Almost instantly, he discovered that hiscreditors were18)11 target: breathing down his neck.18)IP target: demanding repayment.18)IS target: generous and sympathetic.18111, context: Adam enjoyed having a cat in the house. It slept in thebed with him, and he found it comforting to fall asleep whilelistening to it purr and feeling it18)IL target: breathing down his neck.IDIOMS1.APX^ 16319)11 context: Harold loved his new job and was glad to be rid of thepressures of his old one. However, he knew that with the salary che had taken, he would have to19)11 target: tighten his belt.19)IP target: cut his expenses.19)IS target: work extra hours.19)IL context: Paul was pleased with having lost 15 pounds. At first, hthought he might have to buy new clothes, but he discovered that hold ones would be fine if he would just19)IL target: tighten his belt.20)11 context: Chris and his friends were fooling around when theyaccidentally broke his mother's treasured crystal vase. Severaldays passed before his mother noticed, and by then Chris wasrelieved to20)11 target: get it off his chest.20)IP target: get his cuts treated.20)IL context: Elmer was working on his car when the jack slipped,pinning him underneath. He was not seriously injured, but it tookthree rescue workers to20)IL target: get it off his chest.METAPHOR.APX^ 164EXPERIMENT 1 - METAPHORS1)MM context: The committee decided to wait six months longer beforeevaluating Laura's performance because they felt she was still1)MM target: green.1)MP target: inexperienced.1)MS target: recovering.1)ML context: Joanne thought that Ted would notice that she had paintedtheir bedroom a different color, but to him it was still1)ML target: green.2)MM context: A large proportion of Canadians feel that the Free TradeAgreement will mean the end of2)MM target: the Maple Leaf.2)MP target: Canada.2)MS target: Nanaimo Bars.2)ML context: One of my favourite memories of childhood is of sitting outunder the trees in early fall, and marvelling at the range of colourand texture in2)ML target: the maple leaf.3)MM context: Corrine had three classes that morning, with a quiz in oneand a presentation in another. By noon she felt that she definitelyneeded to stop for3)MM target: fuel.3)MP target: food.3)MS target: exercise.3)ML context: Art enjoyed the mental challenge of car rallies even morethan the physical. One of his greatest delights was to keep aconstant record of his car's consumption of3)ML target: fuel.4)MM context: Angela's parents disliked her smoking. They warned her thatbesides being unattractive, cigarette smoking could be a4)MM target: time bomb.4)MP target: latent killer.4)MS target: growth inhibitor.4)ML context: When a child refused to pass his teddy bear through themetal detector at the airport, the official forcibly took the bearaway. He explained that any object had to be considered a potential4)ML target: time bomb.5)MM context: The politician's presentation was smooth, but the facts hepresented were not consistent with what Terry knew to be true. Shesuspected that the data had been5)MM target: cooked.5)MP target: changed.5)MS target: stolen.5)ML context: Marilyn spent the morning in the orchard, picking apples forapplesauce. She didn't step working until evening, when the apples-were—finally peeled, cored, and5)ML target: cooked.METAPHOR.APX^ 1656)MM context: It is important to realize that any crative endeavornecessarily produces frustrations and mistakes. Not even Einsteinideas were all6)MM target: gems.6)MP target: valuable.6)MS target: his own.6)ML context: Buying jewelry today is tricky. With so many syntheticsavailable, the unwary buyer can easily be misled into buyingartificial materials, believing that they are6)ML target: gems.7)MM context: Bert warned his new neighbors to stay away from Dr. McNeil.He had been sued twice for malpractice, and had the reputation intown of being7)MM target: a butcher.7)MP target: incompetent.7)MS target: a peeping tom.7)ML context: The Hills were delighted to find a tiny shopping districtnear their new house which featured a produce market, a coffee shopa baker, and7)ML target: a butcher.8)MM context: It was a perfect summer evening. The lake was deserted butwe could hear the loons calling. The last rays of the sun made thelake a shimmering8)MM target: sheet of foil.8)MP target: sheet of light.8)MS target: mass of fish.8)ML context: Cleaning up after the party was as big a job as preparingfor it. Owen vacuumed the living room, while Liz wrapped up theleftovers in a8)ML target: sheet of foil.9)MM context: The babysitter grew increasingly annoyed as the crashingsounds and yelling increased. Finally she decided it was time todiscipline the9)MM target: troops.9)MP target: children.9)MS target: dogs.9)ML context: The soldiers had been in the jungle for six months and weresuffering from disease and fatigue. The general decided it was timeto replace them with fresh9)ML target: troops.10)MM context: Fluffy was older than Brenda's children, and they regardedher as a large, dignified cat, while Brenda could still remember heras a tiny striped10)MM target: ball.10)MP target: kitten.10)MS target: tiger.___111)MT,centext: Tire game was over almost before it had started becauseErnie hit a grand slam home run, and afterwards nobody could findthe10)ML target: ball.METAPHOR.APX^ 16611)MM context: Karen came into the office wearing a new outfit almostevery week. She expected us to compliment her, but everyone gottired of her continual11)MM target: fishing.11)MP target: hinting.11)MS target: role-playing.11)ML context: The highlight of my summers was always Uncle Barney'svisit, when he would take my brothers and me up north for a week tido some11)ML target: fishing.12)MS context: After Kenley lost his job, he decided that he was just ashappy living on welfare. His old friends knew that his indifferencwas only a12)MM target: facade.12)MP target: pretense.12)MS target: symptom.12)ML context: The architects decided that the building would have to berebuilt, but that they could construct a whole new building andmaintain the old12)ML target: facade.13)MM context: When Peter married Ellen, it was because she had enoughmoney to provide him with whatever he wanted. Ellen had no ideathat he was so13)MM target: calculating.13)MP target: sly.13)MS target: clever.13)ML context: James always dreaded his monthly bookkeeping. He wasnever good at math, and his balances often came out wrong even afterhours of13)ML target: calculating.14)MM context: The police were reluctant to call off their search for themissing child until they were sure that the area had been completely14)MM target: scoured.14)MP target: searched.14)MS target: quarantined.14)ML context: Tina was the best housekeeper we ever had. She took pridein her work, and liked to see floors waxed, silver polished, andsinks14)ML target: scoured.15)MM context: It was bad enough to lose $135, but Chuck also had toreplace all his credit cards after he discovered his wallet had been15)MM target: ripped off.15)MP target: stolen.15)MS target: burned.15)ML context: Actors often wear clothes that are held - together withVelcro- for those moments when the script calls for their costumesto be15)ML target: ripped off.METAPHOR.APX^ 16716)MM context: It was time to call a strike. None of the changes thatmanagement proposed had been made, and the workers were starting imake16)MM target: noises.16)MP target: accusations.16)MS target: mistakes.16)ML context: The mechanics at Mike's Garage amused one another byimitating the sounds customers would make as they tried to describtheir cars's odd16)ML target: noises.17)MM context: After the disasterous 1929 stock market crash, a lot offormerly prosperous financiers saw their future prospects suddenlyturn17)MM target: black.17)MP target: bleak.17)MS target: illegal.17)ML context: Fashion is fickle. At different times, both sophisticatedsociety and rebellious youth have focused their tastes around thecolor17)ML target: black.18)MM context: The studio audiences seemed to love the new game show host,but the critics and a number of the people who watched the show feltthat he was too18)MM context: plastic.18)MP target: artificial.18)MS target: short.18)ML context: One of the biggest problems faced by currentenvironmentally conscious governments is what can be done to stopthe production of throwaway18)ML target: plastic.19)MM context: Alice's playmates thought it was funny to pull the beetle'slegs off one by one and watch it struggle to walk, but Alicethought it was19)MM target: sick.19)MP target: disgusting.19)MS target: inspiring.19)ML context: It is often a difficult challenge for working couples torearrange their day in order to get adequate home care when theirchild is19)ML target: sick.20)MM context: There has been a lot of inefficiency in our administrationof funds. Our new policy is to manage all our funding under asingle20)MM target: umbrella.20)MP target: agency.20)MS target: dictator.contemn It was a grey, rainy day and Harriet dozed on the bus untilshe reached her stop. After she had gotten off, she realized thatshe had left behind her20)ML target: umbrella.SENTENCE COMPREHENSION^ 168Odie Geiger - Psychfest, 1990name:^major:class credit to:(instructor)^^Is^English your most familiar language?^Please answer the following questions with "T" if the meaning onthe question is the same as one of the items you were given toread, or "F" if its meaning is different. Some of the items inthe experiment were presented more than once with differentendings; as long as the question is true of ONE of those youremember, respond "T". Once again, there are no trick questionshere. For example, names are not switched between items, andincidents have not been combined.1.^ Actors wear clothing that is made to come aparteasily.2.^ The kids built a miniature farm while theirbabysitter was on the telephone.3.^ Alice's playmates tried to get beetles to fight overbits of grain.^4.^ The house Linda inherited smelled of rats.5.  Mary joined the school newspaper as a way of makingnew friends.6.^ The housebound wedding party ended up remodelling thegarage.7.^ In the winter sun, the lake looked like a whitetablecloth.8.^ The Free Trade agreement could mean the end ofNanaimo Bars.9.^ The politicians presentation was smooth but his factsseemed suspicious.10.^ Barry was late for work because he got a trafficticket for speeding.11.^ Charles decided to use the same strategy as Neil, andleave the work to someone else.12.^ Max bought his house, car,*and boat with money heinherited.13.^ George and Martha both had faults, but they had agood relationship in spite of them.14.^ Kenley's friends felt that his contentment withwelfare was a symptom of his distress.15.^ Brenda's children loved watching the cat chase a redball of yarn down the stairs.16.^ The new game show host impressed the critics ashighly talented.17.^ It was felt that funds could be better administeredunder the direction of a single agency.18.^ When John asked about a program of physical fitness,the doctor said he was in fine shape.19.^ Dean apologized to Joan for revealing the se-ret ofher plans for a surprise party.20.^ Black has been the color of dress for both thesophisticated and the rebellious.21.^ The mechanics at Mike's garage amused themselveswith video movies.169^22.^ Her office mates felt that Karen was fishing forcompliments on her new outfits.^23.^ Carmen told patients who were having laboratory workdone to look away and hold their breath.24.^ Ralph joined the young men playing football in thepark.25.^ The police refused to continue their search for themissing child.26.^ Ben regularly lost his money while gambling in Reno.27.^ When the child refused to pas his teddy bear throughthe metal detector, the officers laughed.28.^ When Ellen married Peter, she had no idea that he wasinterested in her money.29.^ Barry became arrogant and disrespectful when hebecame a teenager.30.^ When Ricky went swimming, he was bitten all over byminnows.31.^ All Einstein's ideas were spectacularly successful.32.^ Corrine was hungry after a busy morning of classes.Metaphor Comprehension - 170APPENDIX BSENTENCE COMPREHENSION 3^171Odie GeigerSeptember, 1990This experiment involves the reading and understanding ofshort groups of sentences. On each trial, a few sentences willbe displayed, followed by a word or two to complete theirmeaning. Your task will be to read the sentences carefully, andwhen you are sure that you understand their meaning up to thatpoint, push one of the buttons on the response box. Immediately,the sentences will disappear and the ending will be displayed.This portion of the experiment will be timed. As soon as youunderstand the ending so that the whole presentation makes sense,push the button again. A new set of sentences will then appearon the screen and the procedure will continue through 160 trials.It is important to the success of the experiment that youunderstand the meaning of the entire presentation. In order toensure that you are reading the sentences carefully enough foryour responses to be meaningful, a short quiz will follow theexperiment. This does not mean that the sentences will be trickyor difficult. You will likely understand them as soon as youread them. Nor does it mean that you need to memorize thesentences, the details, or the names of the people who may beincidentally mentioned. It is just a way of keeping you fromresponding automatically as soon as the final words appear on thescreen.Since the initial portion of the presentation is not timed,you can leave it on the screen as long as you like to make surethat its meaning is clear to you. After you press the button todisplay the ending, you should respond as quickly as possiblewith a second press of the button when the meaning of the wholephrase is understood. We will provide a few practice trialsbefore the experiment begins to accustom you to the responseprocedure.You will have eight sets of 20 trials each to complete. Ifyou become tired at any point, please feel free to take a breakbetween sets. You will be given a point of class credit and anexplanation of what the experiment is about when the session isover.Thank you for your interest and participation.Odie GeigerGraduate StudentDEADMET2.APX^ 172EXPERIMENT 2 - DEAD METAPHORS1)DD context: The mountaineers decided that they needed two camps to stoisupplies for their climb. One would be midway up the mountain, andthe other would be at the1)DD target: foot.1)DP target: base.1)DS target: castle.1)DL context: A whole day of Christmas shopping was too much for Jessica.When she got home she discovered that she had developed a hugeblister on her1)DL target: foot.2)DD context: Tim's teachers were impressed with him from the start. He h,the very appealing advantage of being both very personable and very2)DD target: bright.2)DP target: smart.2)DS target: wealthy.2)DL context: After nearly two weeks of rain and overcast skies, Myrtle wishocked to look outside her window and see that the sky was clear anc2)DL target: bright.3)DD context: Steve tried to establish a good relationship with his newmother-in-law, but despite all his efforts, she remained3)DD target: cold.3)DP target: unfriendly.3)DS target: shy.3)DL context: Alana loved the idea of going to study as an exchange studenin Norway. All she knew about it was that it would be3)DL target: cold.4)DD context: The camera crew worked hard continuously from 6 a.m. tilnoon, when the director called for a much needed4)DD target: break.4)DP target: rest.4)DS target: swim.4)DL context: Kenny didn't know the road was under construction until histruck bounced into the air, causing most of the 200 dozen eggs he wascarrying to4)DL target: break.5)DD context: All Roger could talk about was cars and baseball. Sally feltthat she had never before met anyone so5)DD target: shallow.5)DP target: uninformed.5)DS target: masculine.5)DL context: The beach at Lost Lake is perfect for young children. Itextends for meters and yet remains sandy, clear and5)DL target: shallow.DEADMET2.APX^ 1736)DD context: After Kenley lost his job, he decided that he was just ashappy living on welfare. His old friends knew that his indifferencewas only a6)DD target: facade.6)DP target: pretense.6)DS terget: symptom.6)DL context: The architects decided that the building would have to berebuilt, but that the could construct a whole new building andmaintain the lovely old6)DL target: facade.7)DD context: Whenever my problems seem to big to handle, I always knowthat I can turn to my Uncle Dave for7)DD target: support.7)DP target: guidance.7)DS target: laughs.7)DL context: One of the most important considerations for engineers ismaking sure that superstructures have adequate7)DL target: support.8)DD context: Angry parents wanted the government to stop cutbacks toschools, but the Department of Education maintained that their planswere8)DD target: firm.8)DP target: definite.8)DS target: progressive.8)DL context: Cheeses have been developed with flavours varying from mildto tangy, and textures ranging from soft to8)DL target: firm.9)DD context: Jason was a superb interior decorator. It was easy torecognize rooms which had his distinctive9)DD target: touch.9)DP target: style.9)DS target: clutter.9)DL context: Good doctors soon develop a sense of how to diagnose theirpatients' problems. They know just what questions to ask, and whichareas to9)DL target: touch.10)DD context: The United States represents a society in which thecultures of a number of different nations have become10)DD target: interwoven.10)DP target: combined.10)DS target: debased.10)DL context: While travelling through India, Lois bought several metersof beautiful silk fabrics into which strands of metallic threadshad been10)DL target: interwoven.DEADMET2.APX^ 17411)DD context: Whenever Toby gets a new mystery novel by her favouriteauthor, she puts everything else aside because she finds it so11)DD target: absorbing.11)DP target: interesting.11)DS target: difficult.11)DL context: The major oil companies are interested in developing amaterial for cleaning up oil spills that is harmless, yet highly o11)DL target: absorbing.12)DD context: Many experienced cyclists are terrified of riding in SoutlAmerica, where roads are in poor repair and often have inadequate12)DD target: shoulders.12)DP target: edges.12)DS target: restrooms.12)DL context: As Grandma got older, she began to wear more warm clothes,even in the summertime. She never went out without putting a shawlover her12)DL target: shoulders.13)DD context: When Pam told Helen that she had eloped with Max againsther parent's wishes and that they were now married, Helen was13)DD target: shocked.13)DP target: surprised.13)DS target: jealous.13)DL context: It is almost certain that if you touch an electricalappliance while you are in the bathtub you will get13)DL target: shocked.14)DD context: In an antique shop in the Maritimes, Jefferey discovered abeautiful old clock with a handpainted14)DD target: face.14)DP target: dial.14)DS target: barometer.14)DL context: Emily could scarcely believe that an eight year old childwould continually try to sneak off to bed without washing his handsand14)DL target: face.15)DD context: Rex liked the idea of living in New York but he worriedabout his health because the smell of industrial pollution was so15)DD target: strong.15)DP target: intense.15)DS target: disguised.15)DL context: Casey chose his food carefully, exercised regularly, andtook vitamin supplements in order to keep himself healthy and15)DL target: strong.DEADMET2.APX^ 17516)DD context: Although she did not like to admit it, Bianca felt that iimany ways her life and her mother's life were16)DD target: parallel.16)DP target: alike.16)DS target: cursed.16)DL context: It must have been very difficult for the workers who builtthe original Canadian National Railway to keep the tracks exactly16)DL target: parallel.17)DD context: The lawyers had assembled a highly detailed and carefullydocumented case against the suspected drug dealers. The judge feltthe facts were17)DD target: clear.17)DP target: correct.17)DS target: distorted.17)DL context: Bermuda is a great place to scuba dive. There is a hugevariety of exotic tropical fish, and the water is warm and17)DL target: clear.18)DD context: Our neighbors called the police when the 30 or 40 teenagersattending the party at the end of the block suddenly got18)DD target: wild.18)DP target: rowdy.18)DS target: quiet.18)DL context: It is interesting for people to be able to see animals inzoos, but it is not the same experience as seeing the animals whenthey are18)DL target: wild.19)DD context: Nigel couldn't understand how Professor Rogers could get somuch attention for his theory when his reasoning seemed so19)DD target: weak.19)DP target: unconvincing.19)DS target: ethnocentric.19)DL context: The hostages were released after four years in captivity.They showed the effects of inadequate food and excercise, and wereextremely19)DL target: weak.20)DD context: When Peter married Ellen, it was because she had enoughmoney to provide him with whatever he wanted. Ellen had no ideathat he was so20)DD target: calculating.20)DP target: ruthless.20)DS target: clever.20)DL context: James always dreaded his monthly bookkeeping. He was neververy good at math, and his balances often came out wrong even afterhours of20)DL target: calculating.MODMET2.APX^ 176EXPERIMENT 2 - METAPHORS1)MM context: The committee decided to wait six months longer beforeevaluating Laura's performance because they felt she was still1)MM target: green.1)MP target: inexperienced.1)MS target: recovering.1)ML context: Joanne thought that Ted would notice that she had paintedtheir bedroom a different color, but to him it was still1)ML target: green.2)MM context: As soon as our garage sale opened, a tall young man with alot of money appeared. He looked over everything carefully, andbought up all the2)MM target: cream.2)MP target: valuables.2)MS target: pillows.2)ML context: Some people like milk in their tea and some people likelemon, or honey, or sugar, but Marilyn is the only person I know whodrinks her tea with2)ML target: cream.3)MM context: Corrine had three classes that morning, with a quiz in oneand a presentation in another. By noon, she felt that she definitelyneeded to stop for3)MM target: fuel.3)MP target: food.3)MS target: exercise.3)ML context: Art enjoyed the mental challenge of car rallies even morethan the physical. One of his greatest delights was to keep a constanrecord of his car's consumption of3)ML target: fuel.4)MM context: When Joel turned 22 his parents insisted that he find anapartment and get a job. They were afraid that he would spend therest of his life as a4)MM target: parasite.4)MP target: dependent.4)MS target: hermit.4)ML context: Albee's greatest fear about going to Africa to work withstarving people was that he might be exposed to some deadly intestinal4)ML target: parasite.5)MM context: The politician's presentation was smooth, but the facts hepresented were not consistent with what Terry knew to be true. Shesuspected that the data had been5)MM target: cooked.5)MP target: changed.5)MS target: stolen.5)ML context: Marilyn spent the morning in the orchard, pinking apples forapplesawee. She didnTt stop working until evening, when the appleswere finally peeled, cored, and5)ML target: cooked.MODMET2.APX^ 1776)MM context: It is important to realize that any creative endeavournecessarily produces frustrations and mistakes. Not evenEinstein's ideas were all6)MM target: gems.6)MP target: valuable.6)MS target: original.6)ML context: Buying jewelry today is tricky. With so many syntheticsavailable, the unwary buyer can easily be mislead into buyingartificial materials, believing that they are6)ML target: gems.7)MM context: Christy couldn't wait to get to the beaches in Hawaii. Shebought a new bikini, but she was determined to be careful. She knewhow easily she could7)MM target: toast.7)MP target: sunburn.7)MS target: drown.7)ML context: It is really hard to adjust to a full English breakfast ofeggs, bacon, sausage, kippers, tomatoes, and mushrooms, if your usuamorning meal consists of coffee and7)ML target: toast.8)MM context: As the newest teacher on staff, Rebecca willingly acceptedresponsibility for the school's Christmas pageant. By October, theproject was becoming a8)MM target: bear.8)MP target: burden.8)MS target: farce.8)ML context: When my cousins from New York came to visit, the thing theywanted to do most was to camp out in the Rockies so they couldphotograph a live8)ML target: bear.9)MM context: The babysitter grew increasingly annoyed as the crashingsounds and yelling increased. Finally she decided it was time todiscipline the9)MM target: troops.9)MP target: children.9)MS target: dogs.9)ML context: The soldiers had been in the jungle for six months and weresuffering from disease and fatigue. The general decided it was timeto replace them with fresh9)ML target: troops.10)MM context: Don's chemistry class was asked to attend a series oflectures by a Nobel Prize winning French chemist. Don went, butfound the lectures difficult to10)MM target: unpack.10)MP target: understand.10)MS target: translate.context: It was incredible that a full year after they had moved,the Gordon's had boxes in their basement that they still hadn'ttaken the time to10)ML target: unpack.MODMET2.APX^ 17811)MM context: Karen came into the office wearing a new outfit almost evweek. She expected us to compliment her, but everyone got tired ofher continual11)MM target: fishing.11)MP target: hinting.11)MS target: theatrics.11)ML context: The highlight of my summers was always Uncle Barney's vis.when he would take my brothers and me up north for a week to do sox11)ML target: fishing.12)MM context: Andrea tried to enhance her lectures with trips anddemonstrations, worked with parents, and coached volleyball. By thetime she got home each night, she was12)MM target: empty.12)MP target: fatigued.12)MS target: sweaty.12)ML context: Paul hoped he would have a chance to talk to one of thespeakers after the conference. He hurried back to the auditorium,but he found it12)ML target: empty.13)MM context: Ross thought he had a good chance of being the next clubpresident, but from the standpoint of the executive committee, hewasn't even in the13)MM target: picture.13)MP target: candidates.13)MS target: club.13)ML context: My neighbor took up photography after he retired. He isextremely talented, and many people have already asked him if theycould buy a13)ML target: picture.14)MM context: Carmen had always tried to think for herself. After she gottoo drunk at an office party, she wondered how she had been able tofollow the14)MM target: wave.14)MP target: crowd.14)MS target: music.14)ML context: The dangers of an earthquake are even greater for peopleliving along a coastline who also have to worry about the possibilit]of a tidal14)ML target: wave.15)MM context: Jason had produced a demonstration recording when he was ateenager. One day he heard a new group on MTV and realized that hisold lyrics had been15)MM target: ripped off.15)MP target: stolen.15)MS target: drivel.15)ML context: Actors often wear clothes that are held together with Velcrofor those moments when the script calls for their costumes to be15)ML target: ripped off.MODMET2.APX^ 17916)MM context: It was time to call a strike. None of the changes thatmanagement proposed had been made, and the workers were starting timake16)MM target: noises.16)MP target: accusations.16)MS target: mistakes.16)ML context: The mechanics at Mike's Garage amused one another byimitating the sounds customers would make as they tried to describctheir car's odd16)ML targets: noises.17)MM context: After the disastrous 1929 stock market crash, a lot offormerly prosperous financiers saw their future prospects suddenlyturn17)MM target: black.17)MP target: bleak.17)MS target: illegal.17)ML context: Fashion is fickle. At different times, both sophisticatedsociety and rebellious youth have focused their tastes around thecolor17)ML target: black.18)MM context: The studio audiences seemed to love the new game show host,but the critics and a number of people who watched the show felt thzhe was too18)MM target: plastic.18)MP target: artificial.18)MS target: short.18)ML context: One of the biggest problems faced by current environmentallconscious governments is what can be done to stop the production ofthrowaway18)ML target: plastic.19)MM context: Alice's playmates thought it was funny to pull the beetle'slegs off one by one and watch it struggle to walk, but Alice thoughtit was19)MM target: sick.19)MP target: disgusting.19)MS target: inspiring.19)ML context: It is often a difficult challenge for working couples torearrange their day in order to get adequate home care when theirchild is19)ML target: sick.20)MM context: There has been a lot of inefficiency in our administrationof funds. Our new policy is to manage all of our funding under asingle20)MM target: umbrella.20)MP target: agency.20)MS target:  bureaucrat. 20)ML context: It was a grey, rainy day and Harriet dozed on the bus untilshe reached her stop. After she had gotten off, she realized thatshe had left behind her20)ML target: umbrella.EXPERIMENT 2^ 180SENTENCE COMPREHENSIONOdie Geigername:^major:class credit to (instructor):^^Is^English your most familiar language?^Please answer the following questions with a "T" if the meaning ofthe question is the same as that of one of the items you were givento read, or "F" if the meaning is different. Some of the items inthe experiment were presented more than once with differentendings; as long as the question is true of any ONE of those youremember, respond "T". Once again, there are no trick questionshere. For example, names are not switched between items, andincidents have not been combined or altered.1.^ Actors wear clothing that is made to come apart easily.2.  Steve wanted to have a good relationship with his newmother-inlaw.3.^ Alice's playmates tried to get the beetles to fight overbits of grain.4.^ All Alana knew about Norway was that it was a cold country.5.^ Rebecca found that producing the school's Christmas pageantwas a delightful experience.6.^ The neighbors offered to chaparone when they heard thatsome teenagers might be having a party.7.^ The facts the politician presented seemed suspicious.8.^ Kenny was amazed to discover that his truckload of eggesmade the trip from Calgary to Vancouver intact.9.^ Kenley's friends felt that his contentment with welfare wasa symptom of his distress.10.^ Rex couldn't notice any air pollution in New York, so hedidn't worry about it.11.^ It was exhausting for Andrea to give so much time andeffort to her classes.12.^The older grandma got, the more she would bundle up to keepwarm.13.^It was great for Sally to find someone who shared herinterest in baseball.14.^ The new game show host impressed the critics as highlytalented.15.^ Society in the United States represents a blending ofcultures.16.^ Cheeses have been developed with a wide variety of texturesand flavors.17.^ It was felt that funds could be better administered undera single agency.18.^ All Einstein's ideas were spectacularly successful.181^19.^ Helen was surprised to hear that Pam and Max had gottenmarried.20.^ Black has been the color of dress for both thesophisticated and the rebellious.21.^ Cyclists are often surprised to find the highways in SouthAmerica so wide and empty.22.^ The mechanics at Mike's Garage amused themselves with videomovies.23.^When Ellen married Peter, she had no idea that he wasinterested in her money.24.^ Her office mates felt that Karen was always looking forcompliments on her new outfits.25.^ Tim was attractive, but his poverty was a liability hecould not overcome.26.^ Jason was in demand as an interior decorator because healways catered to the tastes of his clients.27.^ Nigel thought Professor Rogers's theory was poorly thoughtout.28.^ The mountaineers set up six camps, one for each day oftheir climb.29.^ The lawyers were able to make a convincing case against thedrug dealers.30.^My Uncle Dave is always ready to help me with my problems.31.^ Corrine was hungry after a busy morning of classes.^32.^ Toby has always loved to read romance novels.Metaphor Comprehension - 182APPENDIX C183SENTENCE COMPREHENSION 6Odie GeigerDecember, 1990This experiment involves the reading and understanding ofshort groups of sentences. On each trial, a few sentences willbe displayed, followed by a word to complete their meaning. Yourtask will be to read the sentences carefully, and when you aresure that you understand their meaning up to that point, push oneof the buttons on the response box. Immediately, the sentenceswill disappear and the ending will be displayed. This portion ofthe experiment will be timed. As soon as you understand theending so that the whole presentation makes sense, push thebutton again. A new set of sentences will then appear on thescreen and the procedure will continue through 200 trials.It is important to the success of the experiment that youunderstand the meaning of the entire presentation. In order toensure that you are reading the sentences carefully enough foryour responses to be meaningful, a short quiz will follow theexperiment. This does not mean that the sentences will be trickyor difficult. You will likely understand them as soon as youread them. Nor does it mean that you need to memorise thesentences, the details, or the names of the people who may beincidentally mentioned. It is just a way of keeping you fromresponding automatically as soon as the final words appear on thescreen.Since the initial portion of the presentation is not timed,you can leave it on the screen as long as you like to make surethat its meaning is clear to you. After you press the button todisplay the ending, you should respond as quickly as possiblewith a second press of the button when the meaning of the wholephrase is understood. We will provide a few practice trialsbefore the experiment begins to accustom you to the responseprocedure.You will have ten sets of 20 trials each to complete. Ifyou become tired at any point, please feel free to take a breakbetween sets. You will be given a point of class credit and anexplanation of what the experiment is about when the session isover.Thank you for your interest and participation.Odie GeigerGraduate StudentDEADUNC3.APX^ 184EXPERIMENT 3 - DEAD METAPHORS1)DD context: The recycling committee quickly got people to donate asteady supply of newspapers to its program. Unfortunately, findinga buyer for the papers was an unexpected1)DD target: bottleneck.1)DP target: difficulty.1)DS target: expense.1)DU target: impediment.1)DL context: My grandmother has some old milk bottles in her basement.They were made to hold unhomogenized milk so that the cream wouldseparate out and collect in the1)DL target: bottleneck.2)DD context: Tim's teachers were impressed with him from the start. Hehad the very appealing advantage of being both very personableand very2)DD target: bright.2)DP target: smart.2)DS target: rich.2)DU target: adroit.2)DL context: After nearly two weeks of rain and overcast skies, Myrtlewas shocked to look outside her window and see that the sky wasclear and2)DL target: bright.3)DD context: Steve tried to establish a good relationship with his newmother-in-law, but despite all his efforts, she remained3)DD target: cold.3)DP target: indifferent.3)DS target: shy.)DU target: aloof.3)DL context: Alana loved the idea of going to study as an exchangestudent in Norway. All she knew about it was that it would be3)DL target: cold.4)DD context: Muriel and Howard were a contented couple most of the time,but when it came to politics, they each had opposing4)DD target: positions.4)DP target: opinions.4)DS target: loyalties.4)DU target: dogmas.4)DL context: Joyce finally felt that she had achieved competence as adancer when she realized how automatically she was able to assumethe basic steps and4)DL target: positions.DEADUNC3.APX^ 1855)DD context: All Roger knew anything about was cars and baseball. Sallyfelt that she had never before met anyone so5)DD target: shallow.5)DP target: superficial.5)DS target: masculine.5)DU target: insipid.5)DL context: The beach at Lost Lake is perfect for young children. Thesand extends for several meters and the water is warm and5)DL target: shallow.6)DD context: Warren realised that his problems were just getting worseas time went on. He wanted to make sense of his life, and spentyears in a desperate search for a simple6)DD target: formula.6)DP target: solution.6)DS target: faith.6)DU target: rubric.6)DL context: When Allison discovered a way to remove grass stains fromclothing, several companies offered her large sums of money if shewould sell them her6)DL target: formula.7)DD context: Whenever my problems seem too big to handle, I always knowthat I can turn to my Uncle Dave for7)DD target: support.7)DP target: comfort.7)DS target: money.7)DU target: succor.7)DL context: One of the most important considerations for engineers ismaking sure that superstructures have adequate7)DL target: support.8)DD context: Angry parents wanted the government to stop cutbacks toschools, but the Department of Education maintained that theirplans were8)DD target: firm.8)DP target: definite.8)DS target: progressive.8)DU target: unequivocal.8)DL context: Cheeses have been developed with flavours varying frommild to tangy, and textures ranging from soft to8)DL target: firm.9)DD context: Tina preferred comedies or musicals to plays with socialor political significance. She hated any form of entertainmentthat was too9)DD target: deep.9)DP target: profound.9)DS target: expensive.9)DU target: abstruse.9)DL context: By the time the drillers actually found enough water tomake pumping it worthwhile, the well was almost 50 meters9)DL target: deep.DEADUNC3.APX^ 18610)DD context: Don's chemistry class was invited to attend a lecture bya Nobel Prize winning French chemist. Don went, but he found thelecture difficult to10)DD target: grasp.10)DP target: understand.10)DS target: translate.10)DU target: comprehend.10)DL context: After Bessie's illness, she was very weak. Her son helped 1installing handrails throughout her house so that she would havesomething to10)DL target: grasp.11)DD context: Whenever Toby gets a new mystery novel by her favouriteauthor, she puts everything else aside because she finds it so11)DD target: absorbing.11)DP target: interesting.11)DS target: difficult.11)DU target: engrossing.11)DL context: The major oil companies are interested in developing amaterial for cleaning up oil spills that is harmless, yet highly oil11)DL target: absorbing.12)DD context: Many experienced cyclists are terrified of riding in SouthAmerica, where roads are in poor repair and often have inadequate12)DD target: shoulders.12)DP target: edges.12)DS target: restaurants.12)DU target: verges.12)DL context: As Grandma got older, she began to wear more warm clothes,even in the summertime. She never went out without putting a shawlover her12)DL target: shoulders.13)DD context: When Pam told Helen that she had eloped with Max againsther parent's wishes and that they were now married, Helen was13)DD target: shocked.13)DP target: surprised.13)DS target: happy.13)DU target: appalled.13)DL context: It is almost certain that if you touch an electricalappliance while you are in the bathtub you will get13)DL target: shocked.14)DD context: It was embarrassing for Jerry to forget his speech.Afterwards, his friends found him in a bar. They had never seenhim so14)DD target: depressed.14)DP target: unhappy.14)DS target: drunk.14) DU -target: disheartened-.14)DL context: Even though we live in a rainy area, in the summer thewater level at our municipal reservoir becomes dangerously14)DL target: depressed.DEADUNC3.APX^ 18715)DD context: Rex liked the idea of living in New York but he worriedabout his health because the smell of industrial pollution was so15)DD target: strong.15)DP target: intense.15)DS target: accepted.15)DU target: pungent.15)DL context: Casey chose his food carefully, exercised regularly, andtook vitamin supplements in order to keep himself healthy and15)DL target: strong.16)DD context: Although she did not like to admit it, Bianca felt thatin many ways her life and her mother's life were16)DD target: parallel.16)DP target: similar.16)DS target: cursed.16)DU target: allied.16)DL context: It must have been very difficult for the workers who builtthe original Canadian National Railway to keep the tracks exactly16)DL target: parallel.17)DD context: The lawyers had assembled a highly detailed and carefullydocumented case against the suspected drug dealers. The judge feltthe facts were17)DD target: clear.17)DP target: obvious.17)DS target: unusual.17)DU target: indisputable.17)DL context: Bermuda is a great place to scuba dive. There is a hugevariety of exotic tropical fish, and the water is warm and17)DL target: clear.18)DD context: Gordon was finally coming home after two years overseas.Sally felt calm, but when she finally spotted him at the airport,her feet began to18)DD target: fly.18)DP target: run.18)DS target: itch.18)DU target: jog.18)DL context: The tiny robin was barely able to stand up when Theresafound him under a tree. It was a proud moment for both of themwhen he finally learned to18)DL target: fly.19)DD context: Nigel couldn't understand how Professor Rogers could getso much attention for his theory when his reasoning seemed so19)DD target: weak.19)DP target: vague.19)DS target: silly.19)DU target: imprec4aa,19)DL context: The hostages were released after four years in captivity.They showed the effects of inadequate food and excercise, and wereextremely19)DL target: weak.DEADUNC3.APX^ 18820)DD context: When Peter married Ellen, it was because she had enoughmoney to provide him with whatever he wanted. Ellen had no ideathat he was so20)DD target: calculating.20)DP target: shrewd.20)DS target: poor.20)DU target: astute.20)DL context: James always dreaded his monthly bookkeeping. He wasnever very good at math, and his balances often came out wrongeven after hours of20)DL target: calculating.MODUNC3.APX^ 189EXPERIMENT 3 - METAPHORS1)MM context: The committee decided to wait six months longer beforeevaluating Laura's performance because they felt she was still1)MM target: green.1)MP target: inexperienced.1)MS target: recovering.1)MU target: unversed.1)ML context: Joanne thought that Ted would notice that she had paintedtheir bedroom a different color, but to him it was still1)ML target: green.2)MM context: David was highly regarded as an office manager. His successcame largely because he recognized that making an organization runsmoothly required a lot of elaborate2)MM target: choreography.2)MP target: planning.2)MS target: rules.2)MU target: designing.2)ML context: The dancers were thrilled to be able to work with someonewho could offer them a program featuring such innovative2)ML target: choreography.3)MM context: Corrine had three classes that morning, with a quiz in oneand a presentation in another. By noon, she felt that she definitelyneeded to stop for3)1414 target: fuel.3)MP target: food.3)MS target: exercise.3)MU target: sustenance.3)ML context: Art enjoyed the mental challenge of car rallies even morethan the physical. One of his greatest delights was to keep a constantrecord of his car's consumption of3)ML target: fuel.4)MM context: When the producer invited Steve for an audition, he knew hiscareer might be determined in those few minutes. He wanted to dowell, but he was aware that he was4)MM target: wobbly.4)MP target: nervous.4)MS target: hiccuping.4)MU target: flustered.4)ML context: We knew how upset Mrs. Murray was when her cat got trappedin a tree, but we didn't expect her to rescue it herself, sinceshe was so old and her ladder was so4)ML target: wobbly.MODUNC3.APX^ 1905)MM context: The United States represents a society in which thecultures of a number of different nations have become5)MM target: interwoven.5)MP target: combined.5)MS target: rebellious.5)MU target: conjoined.5)ML context: While travelling through India, Lois bought several metersof beautiful silk fabrics into which strands of metallic thread hadbeen5)ML target: interwoven.6)MM context: It is important to realize that any creative endeavournecessarily produces frustrations and mistakes. Not even Einstein'sideas were all6)MM target: gems.6)MP target: valuable.6)MS target: original.6)MU target: laudable.6)ML context: Buying jewelry today is tricky. With so many syntheticsavailable, the unwary buyer can easily be mislead into buyingartificial materials, believing that they are6)ML target: gems.7)MM context: Carla told her psychiatrist that a full time job, plus ahusband, two children, and an elderly father at home was too muchfor her to7)MM target: juggle.7)MP target: manage.7)MS target: conceal.7)MU target: superintend.7)ML context: When the stress of schoolwork made him tense, Charlie usedto eat. He decided he needed a new pastime, and found exactly thediversion he needed when he learned to7)ML target: juggle.8)MM context: As the newest teacher on staff, Rebecca willingly acceptedresponsibility for producing the school's Christmas pageant. ByOctober, the project was becoming a8)MM target: bear.8)MP target: burden.8)MS target: farce.8)MU target: tribulation.8)ML context: When my cousins from New York came to visit, the thing theywanted to do most was to camp out in the Rockies so they couldphotograph a live8)ML target: bear.MODUNC3.APX^ 1919)MM context: Walking into Margie's house is a retreat from the hectic,noisy world. Margie is a true friend, someone whose nature isalways pleasant, gentle, and9)MM target: sunny.9)MP target: cheerful.9)MS target: simple.9)MU target: genial.9)ML context: Even though it was quite cold, we had a wonderful weekendof skiing. The snow stopped falling just after we arrived, and theskies became9)ML target: sunny.10)MM context: The politician's presentation was smooth, but the factshe presented were not consistent with what Terry knew to be true.She suspected the data had been10)MM target: cooked.10)MP target: changed.10)MS target: stolen.10)MU target: contrived.10)ML context: Marilyn spent the morning in the orchard, picking applesfor applesauce. She didn't stop working until evening, when theapples were finally peeled and10)ML target: cooked.11)MM context: After his team finished its second consecutive season atthe very bottom of the league, the coach decided that his soccerteam needed some new11)MM target: blood.11)MP target: players.11)MS target: uniforms.11)MU target: athletes.11)ML context: Jerry wanted to be an ambulance attendant and help people.In order to get the job, he had to train himself to be insensitiveto the sight of11)ML target: blood.12)MM context: Andrea tried to teach English, PE, and science to studentswhose lives were only focused on personal concerns. Each night shewent home feeling12)MM target: empty.12)MS target: sweaty.12)MU target: ineffectual.12)MP target: useless.12)ML context: Paul hoped he would have a chance to talk to one of thespeakers after the conference. He hurried back to the auditorium,but he found it12)ML target: empty.MODUNC3.APX^ 19213)MM context: Elsie, Bob, and Fluffy visited Julie at Christmas. When herbrother phoned a week later to say that he and his dog would bearriving soon, Julie knew she couldn't stand another13)MM target: attack.13)MP target: guest.13)MS target: animal.13)MU target: caller.13)ML context: When my father talks about his experiences in the war, Ifeel lucky that he's still alive. He often tells how just sixsoldiers were left to defend a whole village against an enemy13)ML target: attack.14)MM context: Carmen had always tried to think for herself. After shegot too drunk at an office party, she wondered how she had beenable to follow the14)MM target: wave.14)MP target: crowd.14)MS target: tune.14)MU target: throng.14)ML context: The dangers of an earthquake are even greater for peopleliving along a coastline who also have to worry about thepossibility of a tidal14)ML target: wave.15)MM context: Edmund had written two novels before he turned 30. Laterhe wrote three more, but his editors rejected them all. Theycomplained that his work had become15)MM target: stagnant.15)MP target: dull.15)MS target: childish.15)MU target: hackneyed.15)ML context: It was shocking for Arnie to revisit his old home townafter 20 years away. Ugly townhouses replaced open fields, andthe pond where he often swam was now15)ML target: stagnant.16)MM context: It was time to call a strike. None of the changes thatmanagement proposed had been made, and the workers were starting tomake16)MM target: noises.16)MP target: complaints.16)MS target: mistakes.16)MU target: innuendoes.16)ML context: The mechanics at Mike's Garage amused one another byimitating the sounds customers would make as they tried todescribe their car's odd16)ML targets: noises.MODUNC3.APX^ 19317)MM context: After the disastrous 1929 stock market crash, a lot offormerly prosperous financiers saw their future prospects suddenlyturn17)MM target: black.17)MP target: hopeless.17)MS target: illegal.17)MU target: glum.17)ML context: Fashion is fickle. At different times, both sophisticatedsociety and rebellious youth have focused their tastes around thecolor17)ML target: black.18)MM context: The studio audiences seemed to love the new game show hostbut the critics and a number of people who watched the show feltthat he was too18)MM target: plastic.18)MP target: artificial.18)MS target: short.18)MU target: affected.18)ML context: One of the biggest problems faced by currentenvironmentally conscious governments is what can be done to stopthe production of throwaway18)ML target: plastic.19)MM context: Alice's playmates thought it was funny to pull thebeetle's legs off one by one and watch it struggle to walk, butAlice thought it was19)MM target: sick.19)MP target: horrible.19)MS target: inspiring.19)MU target: macabre.19)ML context: It is often a difficult challenge for working couples torearrange their day in order to get adequate home care when theirchild is19)ML target: sick.20)MM context: There has been a lot of inefficiency in our administrationof funds. Our new policy is to manage all of our funding under asingle20)MM target: umbrella.20)14P target: agency.20)MS target: bureaucrat.20)MU target: federation.20)ML context: It was a grey, rainy day and Harriet dozed on the busuntil she reached her stop. After she had gotten off, sherealized that she had forgotten her20)ML target: umbrella.194EXPERIMENT 3SENTENCE COMPREHENSION 6Odie Geigername:^major:class credit to (instructor):^Is English your most familiar language?^Please answer the following questions with a "T" if the meaning ofthe question is the same as that of one of the items you were givento read, or "F" if its meaning is different. Some of the items inthe experiment were presented more than once with differentendings; as long as the question s true of any ONE of those youremember, respond "T". ONce again, there are no trick questionshere. For example, names are not switched between items, andincidents have not been combined or altered.1.^Laura's committee recommended an immediate promotion forher.^2.^ Tim was smart, but his rudeness tended to alienate histeachers.3.^ Alice's playmates tried to get the beetles to fight overbits of grain.4.^ All Alana knew about Norway was that it was a cold country.5.^ Rebecca found that producing the school's Christmas pageantwas a delightful experience.6.^ It was David's policy to make plans that would keep hisoffice functioning smoothly.7.^ Terry didn't trust the facts the politician prsented.8.  Don provided his English speaking friends with a completetranscript of the French professor's talk.9.^ The recycling committee had trouble finding buyers fortheir collection of old newspapers.10.^ Rex couldn't notice any air pollution in New YOrk, so hedidn't worry about it.11.^ It was exhausting for Andrea to give so much time andeffort to her classes.12.^ The older grandma got, the more she would bundle up to keepwarm.13.^ It was great for Sally to find someone who shared herinterest in cars and baseball.14.^ The new game show host impressed the critics as highlytalented.15.^ Society in the United States represents a blending ofcultures.16.^ Cheeses have been developed with a wide variety of texturesand flavors.17.^ It was felt that funds could be better administered undera single agency.195^18.^ All Einstein's ideas were spectacularly successful.19.^ Helen was surprised to hear that Pam and Max had gottenmarried.20.^ Black has been the color of dress for both thesophisticated and the rebellious.21.^ Cyclists are often surprised to find the highways in SouthAmerica so wide and empty.22.^ The mechanics at Mike's Garage amused themselves with videomovies.23.^ When Ellen married Peter, she had no idea that he wasinterested in her money.24.^ The coach was looking forward to leading his team intotheir third successful year.25.^ Julie had always looked forward to the visits by her familyand friends.26.^ Tina's taste in drama tended towards Shakespeare and otherdramatists of the Renaissance.27.^ Nigel thought Professor Roger's theory was poorly thoughtout.28.^Jerry's friends rewarded his performance with a triumphanttoast.29.^ The lawyers were able to make a convincing case against thedrug dealers.30.^My Uncle Dave is always ready to help me with my problems.31.^ Corrine was hungry after a busy morning of classes.32.^ Toby has always preferred to read science fiction novels.Metaphor Comprehension - 196APPENDIX D197SENTENCE COMPLETION EXPERIMENT - OCTOBER. 1991 In this experiment, you will be given twenty different shortcontexts, usually one or two sentences long. The last word ofthe last sentence in each case will be missing. For eachcontext, you are to list as many single, English words as youcan think of that will complete the meaning of the sentence.TRY TO BE AS FLEXIBLE AND AS CREATIVE AS POSSIBLE. Nouns,verbs, adverbs, adjectives, metaphors, unusual or rare wordsare all acceptable, as long as (1) you have heard, read orused the word sometime before, (2) you know that you areusing it correctly, and (3) the completed sentence makessense to you. As an example, look over the following sample:Corrine had three classesand a presentation indefinitely needed tofooda.^that morning,^with a quiz in one^another. By noon,^she felt that shestop for sustenanceb. fuel g.exercisec. recharging h. coffeed. sandwiches a nape. lunch j. prayerYou will be allowed 2 minutes to come up with as manyalternatives as you can. After this time, a beeper willsound to indicate that you are to go on to the next sentence.The screen beside you will inform you of which questionnumber you should be on. Since this is an exercise increativity, we would like to have you try to approach thesentences in as many ways as possible. Articles before words(a, the) are acceptable, but except for these, only singleword completions are allowed. Try not to fill up spaces withsynonyms of a single idea (coffee, tea, beer, juice, etc.)unless they express the idea in a different way (potables).If you think of an appropriate word for an earlier questionafter you have moved to a later one, write it down in themargin. After you finish, you can go back and add it to thelist in the spaces labelled with an asterisk. Try as hard asyou can to generate new ideas, but don't be disappointed ityou can't fill all the allotted spaces in the time provided.This experiment should take about an hour, and you willreceive one point of class credit. After the experiment isover, its purpose will be explained to you. Thank you foryour interest and participation.Gei ger,e^Research Assistant1. The committee decided to wait six months longer beforeevaluating Laura's performance because they felt she wasstill ^ .a.  f.^b.^  g.c. h.d.^  1.^e . ^  J. 2.Tim,s teachers were impressed with him from the start. Hehad the very appealing advantage of being both verypersonable and very ^.a.^  f.^b.  g.c.^  h.^d.  1.e. ^  J. ^it. *.3. David was highly regarded as an office manager. Hissuccess came largely because he recognized that making an'organization run smoothly required a lot of elaboratea.^  f.^b.  9.c.^  h.^d.  1.e.^  J.^*. *.1984. Steve tried to establish a good relationship with his newmother-in-law, but despite all his efforts, she remaineda.^  f.^b.  g.c.^  h.^d.  1.e.^  j.^5. It is important to realize that any creative endeavournecessarily produced frustrations and mistakes. Not evenEinstein's ideas were all ^a.^  f.b.  g.^c.^  h.d.  1.^e.^  j. * 6.All Roger knew about anything about was cars and baseball.Sally felt that she had never before met anyone so^•a.^  f.^b.  g.c.^  h.^d.* .^ *1997. Carla told her psychiatrist that a full time job, plus ahusband, two children, and an elderly father at home was toomuch for her to ^ .a.b.^  g.^c. h.d.^  I.^e. j.8. Warren realised that his problems were just getting worseas time went on. He wanted to make sense of his life, andspent years in a desperate search for a simple^ .a.  f.^b.^  g.c.  h.^d.^  I.e.  i. ^9. The politician's presentation was smooth, but the facts hepresented were not consistent with what Terry knew to betrue. She suspected the data had been ^a.^  f.^b.  g.c.^  h.^d.  I.e.^  J.^20010. Angry parents wanted the government to stop cutbacks toschools, but the Department of Education maintained thattheir plans were ^a.^  f.^b.  g.c.^  h.^d.e.^  j•^11. Andrea tried to teach English, PE, and science tostudents whose lives were only focused on personal concerns.Each night she vent home feeling ^a.^  f.b.  g.^c.^  h.d.e.^  j. ^-12. Tina preferred comedies or musicals to plays with socialor political significance. She hated any form of^•entertainment that was too ^a.^  f.b.  g.^c.^  h.d.  I.^e.^  J.20120213. Elsie, Bob, and Fluffy visited Julie at Christmas. Whenher brother phoned a week later to say that he and his dogwould be arriving soon, Julie knew she couldn't stand anothera.^  f.^b.  g.c.^  h.^d.  i.e.^  j.^14. Don's chemistry class was invited to attend a lecture bya Nobel Prize winning French chemist. Don vent, but he foundthe lecture difficult to ^a.^  f.^b.  g.c.^  h.^d.e.^  j.^15. Edmund had written two novels before he turned 30. Laterhe wrote three more, but his editors rejected them all. Theycomplained that his work had become  a.^  f.^b.  g.c.^  h.^d.  i.e.^  J.^*. *.20316. Whenever Toby gets a new mystery novel by her favouriteauthor, she puts everything else aside because she finds itsoa.^  f.^b.  g.c.^  h.^d.e.^  J.^17. The studio audiences seemed to love the new game showhost, but the critics and a number of people who vatched theshow felt that he was too ^a.^  f.b.  g.^c.^  h.d.e.^  J.^*18. When Pam told Helen that she had eloped vith Max againsther parent's wishes and that they were nov married, Helen•wasa.^  f.^b.  g.c.^  h.^d.e.^  J.^19. Alice's playmates thought it was funny to pull thebeetle's legs off one by one and watch it struggle to walk,but Alice thought it was ^a.^  f.^b.  g.c.^  h.^d.e.^  j.^20. Although she did not like to admit it, Bianca felt thatin many ways her life and her mother's life werea.^  f.^b.  g.c.^  h.^d.  I.e .^2042051.^The^committee^decided^toevaluating Laura's performance wait^six^months^longer^beforebecause they felt she was still1. active [1] 51. researching [1]2. adjusting [1] 52. serious [1]3. afraid [1] 53. shocked [1]4. agitated [1] 54. shy [1]5. [an] amateur [1] 55. sick^[3]6. angry [1] 56. single^[1]7. blossoming [1] 57. training [1]8. changing [1] 58. trying [1]9. [a] convict [1] 59. unadapted [1]10. dependent [1] 60. unable [1]11. depressed [1] 61. anacceptable [1]12. developing [2] 62. uncommitted [1]13. doubtful [1] 63. underfed [1]14. early [1] 64. uneasy [1]15. excellent [1] 65. unfit^[1]16. exercising [1] 66. unfoccused [1]17. fat [1] 67. unprepared [2]18. fearful [1] 68. unqualified [1]19. frightened [1] 69. unready [1]20. growing [2] 70. unsuitable [2]21. hurt [1] 71. unsure [1]22. hurting [1] 72. upset [1]23. illiterate [1] 73. winning [1]24. immature [2] 74. working [1]25. improving [3] 75. young [7]26. inadequate [1]27. inappropriate [1] 76. in school [1]28. incompetent [3] 77. not ready [1]29. incomplete [1] 78. too serious [1]30. inconsistent [1]31. indecisive [1]32. inept [1]33. inexperienced [2]34. injured [2]35. insecure 1[]36. irrational [1]37. [a] klutz^[1]38. learning [4]39. [a] nerd^[1]40. new [2]41. nonconfident [1]42. overweight [1]43. practicing [4]44. pregnant [1]45. preparing [1]46. psychotic [1]47. queasy [1]48. quiet [1]49_ recouperating [1]50. recovering [1]2062. Tim's teachers were impressed with him from the start. He hadthe very appealing advantage of being both very personable and very1. active [1] 50. likeable [1]2. aggressive [1] 51. lovable [1]3. appealing [1] 52. masculine [1]4. artistic [1] 53. mature [1]5. articulative^(articulate) 54. neat [1][1] 55. nice [2]6. assertive [1] 56. open [1]7. athletic [5] 57. open-minded [1]8. attractive [3] 58. optimistic [1]9. brave [1] 59. organized [1]10. bright [3] 60. original [1]11. businesslike [1] 61. outgoing [2]12. capable [1] 62. outspoken [2]13. charismatic [1] 63. patient [1]14. charming [1] 64. physical [1]15. committed [1] 65. polite [4]16. conscientious [1] 66. quiet [2]17. considerate [2] 67. respectable [1]18. cool [1] 68. responsible [1]19. cooperative [1] 69. rich [1]20. courageous [1] 70. sexy [1]21. creative [4] 71. smart [2]22. curious [1] 72. sociable [1]23. cute [2] 73. soft-spoken [1]24. dedicated [1] 74. studious [3]25. dependable [1] 75. sympathetic [1]26. direct [1] 76. talkative [2]27. eager [1] 77. tall^[1]28. efficient [2] 78. tardy [1]29. energetic [1] 79. uncompromising [1]30. entertaining [1] 80. understanding [2]31. enthusiastic [1] 81. vocal [1]32. erudite [1] 82. well-dressed [1]33. friendly [2] 83. witty [2]34. funny [1] 84. young [1]35. gentle [1]36. good-looking [1]37. handsome [2]38. hard-working [1]39. helpful [4]40. honest [4]41. humourous [2]42. independent [1]43. informative [1]44. ingenious [1]45. intelligible [1]46. intelligent [7]47. interested [1]4R. kind [1]49. knowledgeable [12073. David was highly regarded ascame largely because he recognizedsmoothly required a lot of elaboratean office manager.^His successthat making an organization run1. administrating [1] 51. orders [1]2. advertising [1] 52. organization [5]3. advisors [1] 53. orgies [1]4. architecture [1] 54. parties [1]5. cases^[1] 55. partners [1]6. changes [1] 56. patience [1]7. clothes [1] 57. pens^[1]8. clothing [1] 58. people [1]9.commands [1] 59. planning [11]10. committees [1] 60. plans [2]11. communication [1] 61. preaching [1]12. companions [1] 62. precision [1]13. computers [1] 63. presentations [1]14. concentration [1] 64. processes [1]15. confrontations [1] 65. programming [1]16. connections [2] 66. promotion [1]17. corrections [1] 67. relationships [1]18. creativity [1] 68. reports [1]19. decisions [1] 69. research [1]20. decorations [1] 70. schedules [1]21. dedication [1] 71. sense [1]22. delegation [1] 72. set-ups [1]23. detailing [1] 73. skill^[3]24. details [1] 74. spelling [1]25. effort [1] 75. spirit [1]26. energy [1] 76. stamina [1]27. enthusiasm [1] 77. strategies [1]28. equipment [1] 78. studying [1]29. expenditures [1] 79. talent [1]30. experiences [1] 80. teamwork [2]31. flattery [1] 81. thinking [2]32. forethought [1] 82. thought [2]33. fortunetelling [1] 83. time^[1]34. friends [1] 84. timing [1]35. furniture [1] 85. timetables [1]36. goals [1] 86. training [1]37. graphs [1] 87. troubleshooting [1]38. guesswork [1] 88. understanding [2]39. ideas^[4] 89. work [4]40. ingratiation [1] 90. workers [2]41. judgements [1]42. juggling [1]43. kindness [1]44. listening [1]45. machinery [1]46. machines [1]47. management [1]48. mediation [1]49_ meetings [2]50. networking [2]2084. Stever tried to establish a good relationship with his newmother-in-law, but despite all his efforts, she remained1. abusive [1]2. accusing [1]3. alcoholic [1]4. alienated [1]5. alive [1]6. aloof [7]7. ambivalent [1]8. angry [2]9. argumentative [1]10. bitter [2]11. brutal [1]12. cold [6]13. cold-eyed [1]14. condescending [1]15. cool [1]16. despicable [1]17. disagreeable [1]18. disapproving [1]19. discontent [1]20. disinterested [2]21. distant [5]22. distrustful [1]23. embarassing [1]24. [an] enemy [1]25. foreign [1]26. frightening [1]27. [a] gorilla [1]28. hateful [1]29. hostile [1]30. ignorant [1]31. impassive [2]32. indifferent [2]33. interfering [1]34. intolerable [1]35. irresponsible [1]36. mean [1]37. obnoxious [1]38. [an] outsider [1]39. pessimistic [2]40. picky [1]41. psychotic [1]42. quiet [3]43. rejecting [1]44. removed [1]45. reserved [1]46. rude [1]47. reclusive [1]48. resentful [3]49. sad [1]50_ sadistic [a] 51. selfish [1]52. self-righteous [1]53. shy [1]54. silent [1]55. similarly [1]56. sleazy [1]57. [a] snob [1]58. snobby [1]59. speechless [1]60. spiteful [1]671. stubborn [2]62. suspicious [1]63. terrible [1]64. unaccepting [1]65. unamused [1]66. unaffected [1]67. unbearable [1]68. uncooperative [1]69. unconscious [1]70. unfair [1]71. unforgiving [1]72. unfriendly [6]73. unkind [1]74. unpalatable [1]75. unpersonable [1]76. unpleasant [1]77. unpleased [1]78. unreachable [1]79. unreasonable [1]80. unsatisfied [2]81. unsociable [1]82. unsympathetic [1]83. untouchable [1]84. unwilling [1]85. upset [2]86. [a] witch [1]87. intalkative [1]2095.1.2.3.4.5.6.7.8.9.10.11.12.13.14.15.16.17.18.19.20.21.22.23.24.25.26.27.28.29.30.31.32.33.34.35.36.37.38.39.40.41.42.43.44.45.46.47.48.49.necessarilyEinstein'sIt^is^important^to^realize^that^any^creative^endeavor^produces^frustrations^and^mistakes.^Notideas were all even[1]acceptable [2]accepted [3]adequate [1]appreciated [1]art [1]believable [1]brilliant [2]Christian [1]clear [2]competent [1]complete [1]comprehensive [1]concise [1]consistent [1]correct [12]creative [1]easy [2]fantastic [1]foolproof [1]fulfilled [1]fulfilling [1]genius [1]god-given [1)good [2]gorgeous [1]great [1]•his^[1]honoured [1]illuminating [1]imaginative [1]important [1]ingenious [2]intelligent [1]interesting [1]jokes^[1]justifiable [1]life-affirming [1]logical [1]normal [1]noteworthy [1]original [2]perfect [6]plausible [3]popular [1]productive [1]proved [1]proven [1]rational [1]rpadablem [1]51.52.53.54.55.56.57.58.59.60.61.62.63.64.65.66.67.68.69.reflective [1]responsible [1]right [3]sane^[2]sensible [1]specific [1]straightforward [2]sound [1]successful [1]true [1]truth [1]understandable [3]useful [3]valid [1]well-founded [1]well-liked [1]wise [1]workable [1]great achievements50. realized [1]2106. All Roger knew anything about was cars and baseball. Sally feltshe had never before met anyone so1.2.3.4.5.6.amazing [1]anal-retentive [1]annoying [1]athletic [1]attached [1]attractive [1]52.53.54.55.56.57.outstanding [1]primitive [1]reclusive [1]rednecked [1]remarkable [1]repetitive [1]7. backward [1] 58. rugged [1]8. boring [8] 59. selfish [1]9. chauvanistic [2] 60. shallow [1]10. childish [1] 61. similar [1]11. compatible [1] 62. simple [2]12. confined [1] 63. single-minded [1]13. crazy [1] 64. smart [1]14. dedicated [1] 65. special [1]15. determined [1] 66. specialized [3]16. distant [1] 67. splendid [1]17. dull^[3] 68. sporty [1]18. egocentrical [1] 69. square [1]19. engrossed [1] 70. static [1]20. entangled [1] 71. stereotypical [2]21. entertaining [2] 72. stupid [7]22. enveloped [1] 73. submerged [1]23. evil^[1] 74. sure [1]24. exciting [2] 75. tall^[1]25. fixated [1] 76. typical [1]26. focused [3] 77. unchallenging [1]27. gorgeous [1] 78. uneducated [3]28. great [1] 79. unimaginative [1]29. grotesque [1] 80. inintelligent [1]30. homey [1] 81. uninteresting [2]31. ignorant [3] 82. unread [1]32. immature [3] 83. useless [1]33. incompatible [1] 84. wonderful [1]34. incompetent [1]35. incredible [1]36. inexperienced [1] 85. carried away [1]37. intense [1]38. interesting [3]39. involved [1]40. irresponsible [1]41. knowledgeable [1]42. lazy [1]43. limited [1]44. loyal^[1]45. masculine [1]46. matched [1]47. mechanical [1]48. narrow-minded [5]49. obsessed [3]50. obsellekte^[1]51. old^[1]2117. Carla told her psychiatrist that a full-time job, plus a husbnd,two children, and an elderly father at home was too much for her to1. abandon [1]2. absorb [1]3. accept [6]4. analyze [1]5. appreciate [2]6. bear [1]7. believe [1]8. clean [1]9. confront [1]10. control [1]11. cope [3]12. do [3]13. envision [1]14. exercise [1]15. expect [1]16. experience [1]17. explain [1]18. feed [1]19. finance [1]20. grasp [1]21. handle [14]22. have [1]23. imagine [1]24. integrate [1]25. juggle [1]26. keep [2]27. leave [1]28. lose [1]29. love [3]30. maintain [1]31. manage [2]32. neglect [1]33. ponder [1]34. relax [2]35. separate [1]36. share [1]37. smile [1]38. stand [2]39. take [3]40. talk [1]41. tolerate [1]42. understand [1]49. live with [1]50. look after [1]51. stress over [1]52. work with [1]43. attend to [1]44. care about [1]45. cook for [1]46. cooperate with [1]47. concentrate on [1]48. listen to [1]2128. Warren realized that his problemstime went on.^Wh wanted to make some senseyears in a desperate search for a simplewere just getting worse asof his life, and spent1. answer [8] 51. vacation [1]2. belief^[1] 52. village [1]3. car [1] 53. wife^[1]4. cereal^[1]5. clue [1]6. companion [2]7. cure [1]8. deity [1]9. drug [1]10. end [2]11. escape [1]12. experiment [1]13. explanation [1]14. friend [1]15. girl^[3]16. goal [4]17. hobby [1]18. home [1]19. house [1]20. ideal^[1]21. incentive [1]22. island [1]23. job^[7]24. key [1]25. life^[3]26. lifestyle [4]27. love [1]28. meaning [2]29. mentor [1]30. path [2]31. pet [1]32. problem [1]33. purpose [2]34. question [2]35. relationship [3]36. religion [3]37. rule [1]38. satisfaction [1]39. sleep [1]40. solution [12]41. suicide [1]42. target [1]43. task [1]44. technique [1]45. time [1]46. toothpaste [1]47. town [1]48. treasure [1]49. truth [3]50. understanding [1]213^9.^The politician'spresented were not consistentsuspected the data had1. added [1]2. adjusted [1]3. altered [4]4. changed [2]5. concocted [1]6. conjured [1]7. contrived [1)8. controlled [1]9. copied [1]10. corrected [1]11. created [2]12. creative [1]13. deleted [2]14. discovered [1]15. doctored [1]16. engineered [1]17. exaggerated [1]18. fabricated [2]19. fake [2]20. faked [1]21. false^[3]22. falsified [5]23. faulty^[1]24. fictitious [1]25. fixed [1]26. forged [2]27. forgotten [1]28. fudged [1]29. half-baked [1]30. hallucinated [1]31. imagined [2]32. imaginary [1]33. inadequate [1]34. inconsistent [1]35. incorrect [2]36. invented [1]37. inverted [1]38. lies^[1]39. limited [1]40. made-up [2]41. manipulated [2]42. miscalculated [1]43. misinterpreted [1]44. misrepresented [1]45. misunderstood [1]46. mixed [1]47. outdated [1]48. overoptimistic [1]49. plagerizcd [1]presentation was smooth,^butwith what Terry knew tobeenthe facts hebe true.^She[1]51.52.53.54.55.56.57.58.59.60.61.62.63.64.65.66.67.random [1]recycled [1]removed [1]selected [1]skewed [1]stacked [1]stolen [1]switched [1]tampered [4]tainted [1]unfounded [1]unproven [1]unreliable [1]untrue [2]worse [1]wrong [4]tampered with50. questionable [1]21410. Angry parents wanted the government to stop cutbacks toschools, but the Department of Education maintained that theirplans were 1. acceptable [1]2. adequate [2]3. appropriate [1]4. approved [1]5. bad[1]6. [the] best [1]7. changing [1]8. continuing [1]9. correct [4]10. critical [1]11. decided [2]12. definite [1]13. demanded [1]14. different [1)15. established [1]16. excellent [1]17. fair [1]18. final (2)19. finalized [1]20. finished [1]21. firm [2]22. fixed [1]23. foolproof [1]24. good [2]25. great [1]26. ignored [1]27. immutable [1]28. implemented [1]29. important [1]30. indifferent [1]31. inevitable [1]32. inflexible (1)33. intelligent [1]34. irreversable [1]35. justifiable [1]36. justified [1]37. law [1]38. long-range [1]39. made [2]40. minimal [1]41. necessary [3]42. normal [1)43. official [1]44. overdue [1]45. permanent [1]46. positive [1]47. practical [1]48. proceeding [1]51. reasonable [3]52. responsible [1]53. responsive [1]54. right [3]55. righteous [1]56. rigid [1]57. settled [1]58. silly [1]59. substantial [1]60. superior [1]61. supportive [1]62. transferred [1]63. unchangeable [2]64. unstoppable [1]65. untransferable [1]66. useful [1]67. useless [1]68. wanted [1]69. well-considered [1]70. wise [1]71. workable [1]72. working [1]73. going ahead [1]74. in action [1]75. in progress [1]50. proven [1]21511.whosewent1.2.3.4.5.6.7.8.9.10.11.12.13.14.15.16.17.18.19.20.21.22.23.24.25.26.27.28.29.30.31.32.33.34.35.36.37.38.39.40.41.42.43.44.45.46.47.48.Andrea tried to teach English,^P.E.lives were only focused on personalhome feelingand science to studentsconcerns.^Each night shenice^[1]nothing [1]old^[1]optimistic [1]overworked [1]pleased [1]positive [1]proud [1]provocative [1]redundant [1]refreshed [1]relaxed [1]revived [1]rejected [1]sad^[1]satisfied [6]scared [1]self-conscious [1]self-righteous [1]sick [1]sleepy [1]spiteful [1]sorry [1]stressful [1]stupid [2]superior [1]tense [1]tired [8]uncertain [1]unfulfilled [4]unhappy [2]unrewarded [1]unwanted [1]unworthy [1]upset [1]useful [2]useless [3]vulnerable [1]wasted [1]weird [1]worried [1]worthless [1]worthwhile [1]unaccomplishments [1]abused [1]alone [2]angry [7]anxious [3]astounded [1]bad [1]beseiged [1]boring [1]brilliant [1]carefree [1]concern [1]confused [1]content [1]cranky [1]crazy [1]defeated [1]dejected [1]depressed [6]detached [1]disappointed [1]discouraged [3]disillusioned [1]disturbed [1]drained [3]elated [1]empty [1]energetic [1]enthusiastic [1]excited [1]exasperated [1]exhausted [2]fatigued [1]frustrated [4]gifted [1]good [1]guilty [1]happy [4]helpful [1]helpless [1]hopeful [1]hopeless [2]immature [1]indecisive [1]ineffective [1]inspired [1]intelligent [1]lost [1]mean [1]51.52.53.54.55.56.57.58.59.60.61.62.63.64.65.66.67.68.69.70.71.72.73.74.75.76.77.78.79.80.81.82.83.84.85.86.87.88.89.90.91.92.93.94.49. meaningless [1]50. miserable [1)12.was1.2.3.4.5.6.7.8.9.10.11.12.13.14.15.16.17.18.19.20.21.22.23.24.25.26.27.28.29.30.31.32.33.34.35.36.37.38.39.40.41.42.43.44.45.46.47.48.politicalTina preferred comediessignificance.^Shetoo216or musicals to plays with social orhated any form of entertainment that51. real^[1]52. realistic [6]53. ridiculous [1]54. sarcastic [1]55. serious^[5]56. silly^[1]57. smart [1]58. snobby [1]59. stuffy [1]60. stupid [1]61. sub-standard [1]62. substantial^[1]63. stressing [1]64. structured [1]65. suggestive [1]66. superficial [1]67. symbolic [1]68. tedious^[1]69. thought-provoking [1]70. tiring [1]71. tricky [1]72. true^[1]73. unbelievable [1]74. violent [2]75. working-class [1]76. worthwhile [1]abstract [1]analytical [1]artistic [1]biased [2]bland [1]bored [1]boring [8]challenging [1]complex [1]conscientious [1]conscious [1]deep [2]degrading [1]delighted [1]depressing [3]didactic [2]dramatic [2]educational [1]enlightening [1]factual [1]fixed [1]focused [1]general [1]graphic [1]heavy [3]highbrow [1]humourous [1]imaginative [1]immature [1]implied [1]inactive [1]informative [1]intellectual [4]intelligent [2]intense [1]lifelike [1]light-hearted [1]mass-produced [1]mature [1]meaningful [3]meaningless [1]moral [1]negative [1]organized [1]political [4]pompouds [1]predictable [1]provocative [1]•50. questionable [1]21713. Elsie, Bob and Fluffy visited Juliebrother phoned a week later to say thatarriving soon, Julie knew she couldn'tat Christmas.^When herhe and his dog would bestand another1. allergy [1] .51. preparation [1]2. animal [6] 52. present [1]3. arrival [1] 53. pressure [1]4. body [1] 54. relative [8]5. bomb [2] 55. responsibility [1]6. brother [1] 56. reunion [1]7. canine [1] 57. sandwich [1]8. Christmas [2] 58. shock [1]9. cleaning [1] 59. slob [1]10. clean-up [1] 60. statement [1]11. complaint [1] 61. tackling [1]12. compliment [1] 62. time [1]13. day [3] 63. troublemaker [1]14. delay [1] 64. vacation [2]15. disappointment [2] 65. visit [7]16. disaster [1] 66. visitor [4]17. disruption [1] 67. voice [1]18. disturbance [1] 68. wait [1]19. dog [6] 69. week [2]20. event [1] 70. word [1]21. face^[1] 71. year [1)22. feast [1]23. fight [2]24. furball [1] 72. dog bone [1]25. geek [1]26. gift [2]27. guest [3]28. hamburger [1]29. holiday [4]30. hotdog [1]31. hour [1]32. houseguest [1]33. inconvenience [1]34. individual [1]35. interference [1]36. jerk [1]37. joker [1]38. licking [1]39. male [1]40. minute [3]41. moment [1]42. movie [1]43. mutt [1]44. news [1]45. night [1]46. noise^[1]47. party [2]48. person [6]• P50. photograph [1]21814. Don's chemistry class was invited to attend a lecture by aNobel Prize winning French chemist. Don went, but he found thelecture difficult to1. absorb [2]2. appreciate [1]3. attend [1]4. awaken [1]5. bear [1]6. believe [3]7. comply [1]8. comprehend [9]9. conceptualize [1]10 . copy [1]11. criticize [1]12. enjoy [2]13. explain [1]14. find [1]15. follow [7]16. grasp [1]17. hate [1]18. hear [10]19. hot [1]20. identify [1]21. incoherent [1]22. intake [1]23. interpret [3]24. leave [1]25. like [1]26. listen [1]27. question [1]28. rationalize [1]29. read [1]30. realize [1]31. relate [1]32. scope [1]33. see [3]34. study [1]35. stuffy [1]36. support [1]37. translate [3]38. understand [14]39. visualize [1]40. concentrate on [1]41. get into [1]42. listen to [1]43. relate to [3]44. write down [1]21915. Edmund had written two novels before he turned 30. Later hewrote^three^more,^but^hiscomplained that his work had editorsbecomerejected^them^all.^They1. abusive [1] 51. political^[1]2. aged [1] 52. predictable [2]3. amateurish [1] 53. prejudice [1]4. arrogant [1] 54. realistic [1]5. bad [1] 55. repetitive [3]6. boring [10] 56. repititious [2]7. childish [1] 57. ridiculous [2]8. commercialistic [1] 58. romantic [1]9. common [1] 59. rusty [1]10. confused [1] 60. sadistic [1]11. crass^[1] 61. serious [2]12. dreary [1] 62. short [1]13. drivel [1] 63. similar [1]14. dry [1] 64. simple [1]15. empty [1] 65. sloppy [4]16. fictitious [1] 66. stale [1]17. garbage [1] 67. static [1]18. gibberish [1] 68. strange [1]19. idiotic [1] 69. stupid [1]20. inadequate [1] 70. stylized [1]21. inarticulate [1] 71. tired [1]22. incoherent [2] 72. trite [1]23. incomprehensible [1] 73. unattractive [1]24. inconcise [1] 74. unbelievable [1]25. indecent [1] 75. unenthusiastic [1]26. inefficient [1] 76. unethical [1]27. infantile [1] 77. uneventful [1]28. insignificant [1] 78. unimportant [1]29. insufficient [1] 79. uninteresting [2]30. irrational [1] 80. unorganized [1]32. irrelevant [1] 81. unoriginal [1]32. journalistic [1] 82. unpopular [1]33. lengthy [1] 83. unreadable [1]34. long-winded [1] 84. unsellable [1]35. mainstream [1] 85. unsuitable [1]36. mature [2] 86. worst [1]37. mediocre [1]38. misleading [1]39. moralizing [1]40. obscene [1]41. obvious [1]42. old [1]43. old-fashioned [1]44. outdated [1]45. outrageous [1]46. passe [1]47. pedantic [1]48. pessimistic [1]50. plain [1]220216.^Whenever Toby geta a new mysteryauthor,^she puts everything elsenovel by her favouriteaside because she finds it so50. meaningful [1]51. mysterious [2]52. pointless^[1]53. provocative [1]54. readable^[2]55. real^[1]56. realistic^[1]57. relaxing [3]58. rewarding [1]59. romantic [1]60. satisfying [1]61. soothing [1]62. spiritual^[1]63. stimulating [1]64. suspenseful [2]65. tasteful^[1]66. tempting [1]67. thoughtful [1]68. thrilling [1]69. time-consuming [1]70. tiring [2]71. transient^[1]72. worthwhile [1]1.2.3.4.5.6.7.8.9.10.11.12.13.14.15.16.17.18.19.20.21.22.23.24.25.26.27.28.29.30.31.32.33.34.35.36.37.38.39.40.41.42.43.44.45.46.47.absorbing [4]addicting [1]alluring [1]amusing [1]appealing [2]attracting [2]attractive [1]awesome [1]boring [1]challenging [2]complex [2]cool^[1]creative [1]descriptive [1]different [1]easy [1]educational [1]engaging [1]engrossing [1]enjoyable [1]entertaining [5]exciting [10]exhilarating [1]familiar [1]fantastic [1]fascinating [2]fast-paced [1]frightening [1]fulfilling [1]fun^[1]funny [1]good [1]hilarious [2]humourous [1]hypnotic [1]imaginative [2]important [2]inspirational [1]intellectual [1]intelligent [1]intense [1]interesting [11]intriguing [5]inviting [1]irrelevant [1]irresistible [2]knowledgeable [1]49. lovable [1]22117.theheThe studio audiences seemedcritics and a number ofwas tooto love the new game show host, butpeople who watched the show felt that1. artificial [1] 51. indifferent [1]2. attractive [1] 52. irresponsible [1]3. aggressive [4] 53. irritating [1]4. apologetic [1] 54. laboured [1]5. arrogant [1] 55. lecherous [2]6. bald [1] 56. loud^[3]7. bizarre [1] 57. loud-mouthed [1]8. black [1] 58. low-class [1]9. blunt [1] 59. manipulative [1]10. boisterous [1] 60. mellow [1]11. bored [1] 61. messy [1]12. boring [3] 62. much [1]13. buoyant [1] 63. nice^[1]14. carefree [1] 64. obnoxious [ 3 ]15. chauvanistic [1] 65. old^[4]16. childish [1] 66. outgoing [1]17. common [1] 67. outrageous [1]18. conservative [ 1 ] 68. outspoken [1]19. crass [1] 69. overbearing [1]20. critical [1] 70. personable [1]21. crude [2] 71. personal [1]22. depressing [2] 72. pessimistic [2]23. derogatory [1] 73. potbellied [1]24. difficult [1] 74. predictable [1]25. doubtful [1] 75. quiet [2]26. dowdy [1] 76. raucous [1]27. dull^[2] 77. remote [1]28. enthusiastic [2] 78. repetitive [1]29. excited [1] 79. rude [1]30. extreme [1] 80. ruthless [1]31. fake [4) 81. sadistic [1]32. false^[1] 82. sarcastic [1]33. fast-speaking [1] 83. sensitive [1]34. fat^[5] 84. short [4]35. flashy [2] 85. showy [1]36. fresh [1] 86. shy [2]37. friendly [1] 87. sickly [1]38. funny [3] 88. simple [1]d39. giddy [1] 89. skinny [1]40. green [1] 90. slick [1]41. gruff [1] 91. slim [1]42. hairy [1] 92. slow [2]43. handsome [2] 93. sly^[1]44. happy [3] 94. smiley [1]45. humourless [1] 95. snooty [1]46. humourous [1] 96. stagnative [1]47. idealistic [1] 97. stand-offish [1]48. immature [1] 98. stereotypical [1]49. impatient [1] 99. stiff [1]50. incoherent [1] 100. strange [1]101. stupid [4]102. superficial [1]103. talkative [2]104. tall [3]105. ugly [4]106. unbelievable [1]107. unconventional [1]108. uncorteous [1]109. unkind [1]110. unoriginal [1]111. vague [1]112. violent [1]113. vocal [1]114. vociferous [1]115. weak [1]116. weird [1]117. wild [1]118. young [2]22222318. When Pam told Helen that she had eloped with Max against herparent's wishes and that they were now married, Helen was1. accepting [1]2. afraid [2]3. aghast [1]4. aloof [1]5. amused [2]6. angry [13]7. annoyed [1]8. appalled [1]9. astonished [3]10. astounded [1]11. baffled [1]12. bathing [1]13. bitter [1]14. calm [1]15. comforting [1]16. confused [3]17. concerned [2]18. cooking [1]19. crying [1]20. curious [1]21. depressed [1]22. disappointed [1]23. disapproving [1]24. disgusted [2]25. drunk [1]26. eating [1]27. ecstatic [3]28. elated [2]29. enchanted [1]30. envious [3]31. excited [4]32. flabbergasted [3]33. furious [2]34. glad [2]35. guilty [1]36. happy [8]37. helpful 1[]38. honest [1]39. horrified [1]40. humoured [1]41. hungry [1]42. hurt [1]43. interested [1]44. itchy [1]45. jealous [5]46. joyous [1]47. judgemental [1]48. impressed [1]49. intrigued [1]50. kind [1] 51. mad [1]52. mean [1]53. melancholy [1]54. moralizing [1]55. mortified [1]56. mystified [1]57. negative [1]58. nervous [1]59. optimistic [1]60. outraged [1]61. overwhelmed [1]62. preoccupied [1]63. proud [1]64. relieved [3]65. sad [1]66. scared [2]67. shocked [11]68. sick [1]69. smiling [1]70. sorry [1]71. speechless [3]72. stunned [1]73. supportive [1]74. surprised [7]75. tearful [1]76. terrified [1]77. triumphant [1]78. understanding [2]79. unforgiving [1]80. unsure [1]81. unsupportive [1]82. upset [3]83. violent [1]84. warm [1]85. worried [4]86. [a] wreck [1]22419. Alice's playmates thought it was funny to pull the beetle'slegs off one by one and watch it struggle to walk, but Alicethought it was1. abnormal [1]2. admirable [1]3. appalling [1]4. awful [3]5. bleeding [1]6. boring [6]7. brilliant [1]8. childish [2]9. cruel [13]10. dead [1)11. degrading [1]12. depressing [1]13. disgusting [7]14. disturbing [1]15. [a] dream [1]16. dying [1]17. escaping [1]18. fascinating [2]19. funnier [1)20. funny [1]21. gross [3]22. harmful [1]23. hilarious [2]24. horrible [2]25. horrifying [1]26. hysterical [1]27. illuminating [1)28. immature [3]29. immoral [1]30. inappropriate [1]31. inhuman[e] [6]32. kind [1]33. mean [4]34. meaningless [1]35. monstrous [1]36. metamorphosing [1]37. neat [1]38. nice [1]39. non-human [1]40. normal [1]41. not [1]42. painful [1]43. right [1]44. [a] ritual [1]45. sad [3]46. saddening [1]47. sadistic [3]48. savagely [1]49. sick [2] 50. spell-binding [1)51. [a] sport [1]52. strange [2]63. stupid [5]54. suffereing [1]55. terrible [1]56. tired [1]57. torture [1]58. tyrannical [1]59. unattractive [1]60. unbelievable [1]61. uncaring [1]62. uncivilized [1]63. uneventful [1]64. unfair [1]65. ungodly [1]66. unhumanistic [1]67. unoriginal [1]68. unusual [1)69. vengeful [1]70. violent [2]71. warped [1]72. [a] waste [1]73. weird [1]74. wrong [1]75. a waste of time [1]22520. Although she did not like to admit it, Bianca felt that in manyways her life and her mother's life were^ .1. abnormal [2] 52. limited [1]2. active [1] 53. linked [1]3. associated [1] 54. lively [1]4. beautiful [1] 55. lonely [2]5. [the] best [1] 56. meaningless [1]6. bizarre [1] 57. opposite [1]7. boring [6] 58. over [1]8. bounded [1] 59. overlapping [1]9. calm [1] 60. pathetic [1]10. chaotic [1] 61. parallel^[1]11. colorful^[1] 62. perfect [1]12. connected [2] 63. poor [1]13. contrastive [1] 64. promising [1]14. controlled [1] 65. repeating [1]15. convergent [1] 66. restricted [1]16. creative [1] 67. routine [1]17. dangerous [3] 68. routined [1]18. dependent [1] 69. ruined [1]19. different [9] 70. sad^[3]20. disparate [1] 71. [the] same [3]21. distant [1] 72. satisfactory [1]22. diverging [1] 73. sedentary [1]23. doomed [1] 74. selfish [1]24. endangered [1] 75. separate [2]25. empty [1] 76. separated [1]26. ending [1] 77. short [1]27. entertaining [1] 78. similar^[12]28. equal [1] 79. spoiled [1]29. exciting [2] 80. spontaneous [1]30. extreme [1] 81. steady [1]31. failures^[1] 82. stressful [1]32. frightening [1] 83. superficial^[1]33. frivolous [1] 84. supressed [1]34. fulfilled [1] 85. tedious [1]35. fulfilling [1] 86. terrible [1]36. fun [1] 87. together [1]37. funny [1] 88. typical [1]38. glamourous [1] 89. unalike [1]39. happy [2] 90. unconventional [1]40. hard [1] 91. unfulfilled [1]41. hectic [1] 92. uninteresting [2]42. hopeless [2] 93. unpredictable [1]43. identical^[2] 94. unusual [2]44. incompatible [1] 95. worthless [1]45. independent [1]46. inseparate [1]47. intellectual [1]48. interconnected [1]49. intertwined [1]cO_ intransigent [1]51. intriguing [1)Metaphor Comprehension - 226APPENDIX E227LEXICAL DECISION - APPROPRIATE VS. INAPPROPRIATE - SENLEXJune, 1992This experiment consists of a series of short sentence contexts in whichthe final word of the final sentence is missing. Your task will be to choosewhich of two words completes the intended meaning of the sentence. Aftermaking your choice, we want you to indicate, first for the chosen word andthen for the rejected word, just how appropriate you think these two choicesare to the sentence context. Using a scale like the one below, give a "1" to thewords you judge to be highly appropriate, and a "7" to the ones you believe areabsolutely inappropriate; the ones that make no sense at all. 'Appropriate'means a word that might actually be used to complete the meaning of thesentence context, not just one that is somehow related to the topic. A wordcan be highly appropriate even it is surprising or unfamiliar in terms of theexpectations generated by the sentences. Do not worry if you find yourselfusing some categories more than others, this is not at all unlikely. Just try togive your honest first response to each sentence in terms of its ending.1^2^3^4^5^6^7highly fairly fairly highlyappropriate^appropriate^inappropriate^inappropriateThere will likely be some words in the experiment that are unfamiliar toyou, or you feel you understand only vaguely. This may be enough to guideyour decision as to which word to choose, but leave you unable to assess itsappropriateness. If you feel you are really too unfamiliar with the word tojudge whether it is appropriate or not, please put an "X" in that space. It isimportant not to confuse unfamiliarity with inappropriateness - that is. do notcall a word inappropriate only because you are unfamiliar with it.The sentence contexts will be displayed all at once on a computermonitor. Most are one to two sentences long. You are to read the sentences,and once you have understood their meaning, press the bottom two buttons onthe box as indicated. This will remove the context and present you with twowords, side by side, on the video screen. Read them both, and decide whichof the two is the most appropriate ending for the context. Then think abouthow appropriate you feel it is, and rank it on the paper according to the 7-point scale. Do the same for the less appropriate word on the second line.The top line for each air should alwa s be a lower num Pi 1 • 1 1 " • 11228line. Once you have made your decision and rated both words, press thebutton on the top of the button box corresponding to the word that you haverated first (the more appropriate word). A new context will appear and therating procedure should continue through 200 trials. We want you to respondas accurately as possible, but the decisions should be fairly spontaneous. Nodecision should take you more than a few seconds.You will receive one class credit for about an hour of participation.Thank you for your interest.Odie Geiger,Research AssistantDr. Lawrence Ward,Supervising Professor2a.229LEXICAL DECISION - APPROPRIATE VS. INAPPROPRIATE - SENLEXJune, 1992Please rank the more appropriate of the two words in the "a" space andthe less appropriate in the 'b" space. Give a "1" to the words you judge to behighly appropriate, and a "7" to the ones you believe are absolutelyinappropriate. If you feel you are too unfamiliar with the word to judgewhether it is appropriate or not, please put an "X" in that space. It isimportant not to confuse unfamiliarity with inappropriateness - that is. do notcall a word inappropriate only because you are unfamiliar with it.1^2^3^4^5^6^7highly^fairly fairlyhighlyappropriate^appropriate^inappropriate^inappropriate************************** PRACTICE *****************************la.lb. 5a.5b.2b.3a.3b.4a.4b.***************************** SET 1 *****************************la. 5a. 9a. 13a. 17a.lb. 5b. 9b. 13b. 17b.2a. 6a. 10a. 14a. 18a.2b. 6b. 10b. 14b. 18b.3a. 7a. lla. 15a. 19a.3b. 7b. 11b. 15b. 19b.4a. 8a. 12a. 16a. 20a.4b. 8b. 12b. 16b. 20b.230* * * * * * * * * * * * * * * * * * * * *** ** *** SET 2 *****************************la.^5a.^9a.^13a.^17a.lb. 5b. 9b. 13b. 17b.2a. 6a.^10a.^14a.^18a.2b. 6b. 10b. 14b. 18b.3a.^7a.^lla.^15a.^19a.3b. 7b. 11b. 15b. 19b.4a. 8a.^12a.^16a.^20a.4b. 8b. 12b. 16b. 20b.*** * * * *^* **** * ** *********** SET 3 ******'************************la.^5a.^9a.^13a.^17a.lb. 5b. 9b. 13b. 17b.2a. 6a.^10a.^14a.^18a.2b. 6b. 10b. 14b. 18b.3a. 7a.^lla.^15a.^19a.3b. 7b. 11b. 15b. 19b.4a. 8a.^12a.^16a.^20a.4b. 8b. 12b. 16b. 20b.***************************** SET 4 *****************************la.^5a.^9a.^13a.^17a.lb. 5b. 9b. 13b. 17b.2a. 6a.^10a.^14a.^18a.2b. 6b. 10b. 14b. 18b.3a. 7a.^lla.^15a.^19a.3b. 7b. 11b. 15b. 19b.4a. 8a.^12a.^16a.^20a.4b. 8b. 12b. 16b. 20b.231* * * * * * * * * * * * * * * * * * ********** * SET 5 *****************************la.^5a.^9a.^13a.^17a.lb. 5b. 9b. 13b. 17b.2a. 6a.^10a.^14a.^18a.2b. 6b. 10b. 14b. 18b.3a. 7a.^lla.^15a.^19a.3b. 7b. 11b. 15b. 19b.4a. 8a.^12a.^16a.^20a.4b. 8b. 12b. 16b. 20b.******* ** ** ****** * ********* SET 6 *****************************la.^5a.^9a.^13a.^17a.lb. 5b. 9b. 13b. 17b.2a. 6a.^10a.^14a.^18a.2b. 6b. 10b. 14b. 18b.3a.^7a.^lla.^15a.^19a.3b. 7b. 11b. 15b. 19b.4a. 8a.^12a.^16a.^20a.4b. 8b. 12b. 16b. 20b.***************************** SET 7 *****************************la.^5a.^9a.^13a.^17a.lb. 5b. 9b. 13b. 17b.2a. 6a. ^10a.^14a.^18a.2b. 6b. 10b. 14b. 18b.3a. 7a.^I la.^15a.^19a.3b. 7b. 11b. 15b. 19b.4a. 8a.^12a.^16a.^20a.4b. 8b. 12b. 16b. 20b.232* * * * * * * * * * * * * * * * * * * ********* SET 8 ******************************la.^5a.^9a.^13a.^17a.lb. 5b. 9b. 13b. 17b.2a. 6a.^10a.^14a.^18a.2b. 6b. 10b. 14b. 18b.3a. 7a.^lla.^15a.^19a.3b. 7b. 11b. 15b. 19b.4a. 8a.^12a.^16a.^20a.4b. 8b. 12b. 16b. 20b.**************************** SET 9 *******************************la. 5a. 9a. 13a. 17a.lb. 5b. 9b. 13b. 17b.2a. 6a. 10a. 14a. 18a.2b. 6b. 10b. 14b. 18b.3a. 7a. 11a. 15a. 19a.3b. 7b. 11b. 15b. 19b.4a. 8a. 12a. 16a. 20a.4b. 8b. 12b. 16b. 20b.***************************** SET *****************************la. 5a. 9a. 13a. 17a.lb. 5b. 9b. 13b. 17b.2a. 6a. 10a. 14a. 18a.2b. 6b. 10b. 14b. 18b.3a. 7a. lla. 15a. 19a.3b. 7b. 1 lb. 15b. 19b.4a. 8a. 12a. 16a. 20a.4b. 8b. 12b. 16b. 20b.233LEXICAL DECISION - APPROPRIATE VS. INAPPROPRIATE - SENLEXJune, 1992This experiment consists of a series of short sentence contexts in whichthe final word of the final sentence is missing. Your task will be to choosewhich of two words completes the intended meaning of the sentence, and thento indicate your choice by pressing a button corresponding to the chosen wordas quickly as possible.The sentence contexts will be displayed all at once on a computermonitor. Most are one or two sentences long. You are to read the sentences,and once you have understood their meaning, press the bottom buttons on thebox as indicated. This will remove the context and present you with two words,side by side, on the video screen. Read them both, and then press either theright or left button on the top of the box corresponding to the position of theword. Once you have made your response, a new sentence context willautomatically appear on the screen.As the presentation continues, you will notice that some of the contextsare repeated. However, the endings for these sentences will be different. Itis important to the success of this experiment that you read through everycontext to the point of understanding each time it appears, even if you recallit from a previous display. There will be a short pencil and paper quizfollowing the experiment that you will find easy to answer as long as eachcontext is understood in terms of its endings.If you are ever unfamiliar with the meaning of one or both of the wordspresented, or if you are unable to decide between the two, or if neither of thewords seems to make sense to you in terms of the context, just make a guessand continue on with the experiment. You should try to work at a steadypace, and to be as accurate as possible in making your choice.The entire experiment consists of 200 trials of sentence contexts followedby the quiz, and should take about one hour to complete. You will receiveone hour of class credit for your participation. Thank you for your interest.Odie Geiger,Research AssistantDr. Lawrence Ward,^SupervisingDEADUNC5.APX^ 234EXPERIMENT 5 - DEAD METAPHORSNote: In this appendix, correct targets are listed first, followed bymatched incorrect targets. In the actual experiment, the correct stimuluwas randomly selected to appear on either the right or left side of thescreen.1)DD context: The recycling committee quickly got people to donate asteady supply of newspapers to its program. Unfortunately, findinga buyer for the papers was an unexpected1)DD target: bottleneck^sketchbook1)DP target: difficulty atmosphere1)DS target: expense welcome1)DU target: impediment^hypodermic1)DL context: My grandmother has some old milk bottles in her basement.They were made to hold unhomogenized milk so that the cream wouldseparate out and collect in the1)DL target: bottleneck^paraphrase2)DD context: Tim's teachers were impressed with him from the start. Hehad the very appealing advantage of being both very personableand very2)DD target: bright^battle2)DP target: smart valid2)DS target: rich gain2)DU target: astute^malted2)DL context: After nearly two weeks of rain and overcast skies, Myrtlewas shocked to look outside her window and see that the sky wasclear and2)DL target: bright^weight3)DD context: Steve tried to establish a good relationship with his newmother-in-law, but despite all his efforts, she remained3)DD target: cold^move3)DP target: indifferent^seventeenth3)DS target: shy par3)DU target: aloof^plaid3)DL context: Alana loved the idea of going to study as an exchangestudent in Norway. All she knew about it was that it would be3)DL target: cold^sort4)DD context: Muriel and Howard were a contented couple most of the time,but when it came to politics, they each had opposing4)DD target: positions^elections4)DP target: opinions supplies4)DS target: loyalties footnotes4)DU target: stances^oatmeal4)DL context: Joyce finally felt that she had achieved competence as adancer when she realized how automatically she was able to assume4)DL target: positions^employeesDEADUNC5.APX2355)DD context: All Roger knew anything about was cars and baseball. Sallyfelt that she had never before met anyone so5)DD target: shallow^pasture5)DP target: superficial^microscopic5)DS target: masculine scratched5)DU target: insipid^twirled5)DL context: The beach at Lost Lake is perfect for young children. Thesand extends for several meters and the water is warm and5)DL target: shallow^amended6)DD context: Warren realised that his problems were just getting worseas time went on. He wanted to make sense of his life, and spentyears in a desperate search for a simple6)DD target: formula^tension6)DP target: solution railroad6)DS target: faith staff6)DU target: tenet^snout6)DL context: When Allison discovered a way to remove grass stains fromclothing, several companies offered her large sums of money if shewould sell them her6)DL target: formula^housing7)DD context: Whenever my problems seem too big to handle, I always knowthat I can turn to my Uncle Dave for7)DD target: support^looking7)DP target: comfort regions7)DS target: money^death7)DU target: sustenance^casualties7)DL context: One of the most important considerations for engineers ismaking sure that superstructures have adequate7)DL target: support^picture8)DD context: Angry parents wanted the government to stop cutbacks toschools, but the Department of Education maintained that theirplans were8)DD target: firm^cars8)DP target: definite^swimming8)DS target: progressive respirtory8)DU target: unequivocal chauffeured8)DL context: Cheeses have been developed with flavours varying frommild to tangy, and textures ranging from soft to8)DL target: firm^pool9)DD context: Tina preferred comedies or musicals to plays with socialor political significance. She hated any form of entertainmentthat was too9)DD target: deep^unit9)DP target: profound^released9)DS target: expensive separatedg •9)DL context: By the time the drillers actually found enough water tomake pumping it worthwhile, the well was almost 50 meters9)DL target: deep^jazzDEADUNC5.APX^ 23610)DD context: Don's chemistry class was invited to attend a lecture bya Nobel Prize winning French chemist. Don went, but he found thelecture difficult to10)DD target: grasp^tumor10)DP target: understand^population10)DS target: translate dimension10)DU target: comprehend accelerate10)DL context: After Bessie's illness, she was very weak. Her son helpedinstalling handrails throughout her house so that she would havesomething to10)DL target: grasp^quote11)DD context: Whenever Toby gets a new mystery novel by her favouriteauthor, she puts everything else aside because she finds it so11)DD target: absorbing^shuffling11)DP target: interesting responsible11)DS target: difficult^committee11)DU target: engrossing unsinkable11)DL context: The major oil companies are interested in developing amaterial for cleaning up oil spills that is harmless, yet highly of11)DL target: absorbing^threading12)DD context: Many experienced cyclists are terrified of riding in SouthAmerica, where roads are in poor repair and often have inadequate12)DD target: shoulders^decisions12)DP target: edges^knees12)DS target: restaurants metaphysics12)DU target: verges rivals12)DL context: As Grandma got older, she began to wear more warm clothes,even in the summertime. She never went out without putting a shawlover her12)DL target: shoulders^elections13)DD context: When Pam told Helen that she had eloped with Max againsther parent's wishes and that they were now married, Helen was13)DD target: shocked^blowing13)DP target: surprised organized13)DS target: happy glass13)DU target: appalled^prolific13)DL context: It is almost certain that if you touch an electricalappliance while you are in the bathtub you will get13)DL target: shocked^broader14)DD context: It was embarrassing for Jerry to forget his speech.Afterwards, his friends found him in a bar. They had never seenhim sotarget: depressedtarget: unhappytarget: drunktarget: disheartenedcontext: Even though we live in a rainy area, in the summer thewater level at our municipal reservoir becomes dangerously14)DL target: depressed^respected14)DD14)DP14)DS14)DU14) DLrenderingstainedmagictranslusccntDEADUNC5.APX^ 23715)DD context: Rex liked the idea of living in New York but he worriedabout his health because the smell of industrial pollution was so15)DD target: strong^beyond15)DP target: intense concert15)DS target: accepted opposite15)DU target: pungent^tubular15)DL context: Casey chose his food carefully, exercised regularly, andtook vitamin supplements in order to keep himself healthy and15)DL target: strong^inside16)DD context: Although she did not like to admit it, Bianca felt thatin many ways her life and her mother's life were16)DD target: parallel^regional16)DP target: similar earlier16)DS target: cursed^porous16)DU target: akin surf16)DL context: It must have been very difficult for the workers who builtthe original Canadian National Railway to keep the tracks exactly16)DL target: parallel^enormous17)DD context: The lawyers had assembled a highly detailed and carefullydocumented case against the suspected drug dealers. The judge feltthe facts were17)DD target: clear^north17)DP target: obvious musical17)DS target: unusual highest17)DU target: indisputable^aeronautical17)DL context: Bermuda is a great place to scuba dive. There is a hugevariety of exotic tropical fish, and the water is warm and17)DL target: clear^alone18)DD context: Gordon was finally coming home after two years overseas.Sally felt calm, but when she finally spotted him at the airport,her feet began to18)DD target: fly^pat18)DP target: run cut18)DS target: itch slam18)DU target: jog^rib18)DL context: The tiny robin was barely able to stand up when Theresafound him under a tree. It was a proud moment for both of themwhen he finally learned to18)DL target: fly^bid19)DD context: Nigel couldn't understand how Professor Rogers could getso much attention for his theory when his reasoning seemed so19)DD target: weak^hurt19)DP target: vague brave19)DS target: silly lined19)DU target: imprecise^sleepless19) DL cvuteXL. TLe hustayes were released after four years in captivity.They showed the effects of inadequate food and excercise, and wereextremely19)DL target: weak^bareDEADUNC5.APX^ 23820)DD context: When Peter married Ellen, it was because she had enoughmoney to provide him with whatever he wanted. Ellen had no ideathat he was so20)DD target: calculating^fragmentary20)DP target: shrewd^barred20)DS target: poor grey20)DU target: wily dank20)DL context: James always dreaded his monthly bookkeeping. He wasnever very good at math, and his balances often came out wrongeven after hours of20)DL target: calculating^adolescentsMODUNC5.APX^ 239EXPERIMENT 5 - METAPHORSNote: In this appendix, correct targets are listed first, followed bymatched incorrect targets. In the actual experiment, the correct stimuliwas randomly selected to appear on either the right or left side of thescreen.1)MM context: The committee decided to wait six months longer beforeevaluating Laura's performance because they felt she was still1)MM target: green^seven1)MP target: inexperienced^environmental1)MS target: recovering synonomous1)MU target: unversed^embossed1)ML context: Joanne thought that Ted would notice that she had paintedtheir bedroom a different color, but to him it was still1)ML target: green^teeth2)MM context: David was highly regarded as an office manager. His successcame largely because he recognized that making an organization runsmoothly required a lot of elaborate2)MM target: choreography^astrophysics2)MP target: planning^religion2)MS target: rules youth2)MU target: designing^figurines2)ML context: The dancers were thrilled to be able to work with someonewho could offer them a program featuring such innovative2)ML target: choreography^bewilderment3)MM context: Corrine had three classes that morning, with a quiz in oneand a presentation in another. By noon, she felt that she definitelyneeded to stop for3)MM target: fuel^sums3)MP target: food hair3)MS target: exercise^machines3)MU target: sustenance highlights3)ML context: Art enjoyed the mental challenge of car rallies even morethan the physical. One of his greatest delights was to keep a constalrecord of his car's consumption of3)ML target: fuel^cats4)MM context: When the producer invited Steve for an audition, he knew hi:career might be determined in those few minutes. He wanted to dowell, but he was aware that he was4)MM target: wobbly^clawed4)MP target: nervous biggest4)MS target: hiccuping^garlanded4)MU target: flustered vexatious4)ML context: We knew how upset Mrs. Murray was when her cat got trappedin a tree, but we didn't expect her to rescue it herself, sinceshe was so old and her ladder was so 4)ML target: wobbly^pitiedMODUNC5.APX^ 2405)MM context: The United States represents a society in which thecultures of a number of different nations have become5)MM target: interwoven^morticians5)MP target: combined hydrogen5)MS target: rebellious bequeathed5)MU target: conjoined^subtitled5)ML context: While travelling through India, Lois bought several metersof beautiful silk fabrics into which strands of metallic thread hadbeen5)ML target: interwoven^reconvened6)MM context: It is important to realize that any creative endeavournecessarily produces frustrations and mistakes. Not even Einstein'sideas were all6)MM target: gems^pegs6)MP target: valuable^changing6)MS target: original returned6)MU target: laudable buttoned6)ML context: Buying jewelry today is tricky. With so many syntheticsavailable, the unwary buyer can easily be mislead into buyingartificial materials, believing that they are6)ML target: gems^hubs7)MM context: Carla told her psychiatrist that a full time job, plus ahusband, two children, and an elderly father at home was too muchfor her to7)MM target: juggle^dilute7)MP target: manage bundle7)MS target: conceal breathe7)MU target: superintend^firecracker7)ML context: When the stress of schoolwork made him tense, Charlie usedto eat. He decided he needed a new pastime, and found exactly thediversion he needed when he learned to7)ML target: juggle^staple8)MM context: As the newest teacher on staff, Rebecca willingly acceptedresponsibility for producing the school's Christmas pageant. ByOctober, the project was becoming a8)MM target: bear^veto8)MP target: burden lawyer8)MS target: farce leash8)MU target: tribulation^requisition8)ML context: When my cousins from New York came to visit, the thing theywanted to do most was to camp out in the Rockies so they couldphotograph a live8)ML target: bear^rustMODUNC5.APX^ 2419)MM context: Walking into Margie's house is a retreat from the hectic,noisy world. Margie is a true friend, someone whose nature isalways pleasant, gentle, and9)MM target: sunny^blade9)MP target: cheerful deceased9)MS target: simple higher9)MU target: genial^dotted9)ML context: Even though it was quite cold, we had a wonderful weekendof skiing. The snow stopped falling just after we arrived, and theskies became9)ML target: sunny^aloud10)MM context: The politician's presentation was smooth, but the factshe presented were not consistent with what Terry knew to be true.She suspected the data had been10)MM target: cooked^caring10)MP target: changed talking10)MS target: stolen cosmic10)MU target: contrived^nocturnal10)ML context: Marilyn spent the morning in the orchard, picking applesfor applesauce. She didn't stop working until evening, when theapples were finally peeled and10)ML target: cooked^reared11)MM context: After his team finished its second consecutive season atthe very bottom of the league, the coach decided that his soccerteam needed some new11)MM target: blood^trees11)MP target: players samples11)MS target: uniforms feathers11)MU target: athletes^excerpts11)ML context: Jerry wanted to be an ambulance attendant and help people.In order to get the job, he had to train himself to be insensitiveto the sight of11)ML target: blood^truth12)MM context: Andrea tried to teach English, PE, and science to studentswhose lives were only focused on personal concerns. Each night shewent home feeling12)MM target: empty^thick12)MS target: sweaty wooded12)MU target: ineffectual^aerodynamic12)MP target: useless^flowing12)ML context: Paul hoped he would have a chance to talk to one of thespeakers after the conference. He hurried back to the auditorium,but he found it12)ML target: empty^idealMODUNC5.APX^ 24213)MM context: Elsie, Bob, and Fluffy visited Julie at Christmas. When h(brother phoned a week later to say that he and his dog would bearriving soon, Julie knew she couldn't stand another13)MM target: assault^profile13)MP target: guest stick13)MS target: animal career13)MU target: caller^pillar13)ML context: When my father talks about his experiences in the war, Ifeel lucky that he's still alive. He often tells how just sixsoldiers were left to defend a whole village against an enemy13)ML target: assault^collage14)MM context: Carmen had always tried to think for herself. After shegot too drunk at an office party, she wondered how she had beenable to follow the14)MM target: wave^salt14)MP target: crowd grass14)MS target: tune^doll14)MU target: horde blush14)ML context: The dangers of an earthquake are even greater for peopleliving along a coastline who also have to worry about thepossibility of a tidal14)ML target: wave^bond15)MM context: Edmund had written two novels before he turned 30. Laterhe wrote three more, but his editors rejected them all. Theycomplained that his work had become15)MM target: stagnant^tapestry15)MP target: dull^sand15)MS target: childish tropical15)MU target: hackneyed^laminated15)ML context: It was shocking for Arnie to revisit his old home townafter 20 years away. Ugly townhouses replaced open fields, andthe pond where he often swam was now15)ML target: stagnant^perverse16)MM context: It was time tomanagement proposed hadmake16)MM target: noises16)MP target: complaints16)MS target: mistakes^arteries16)MU target: innuendoes tombstones16)ML context: The mechanics at Mike's Garage amused one another byimitating the sounds customers would make as they tried todescribe their car's odd16)ML targets: noises^stormscall a strike. None of the changes thatbeen made, and the workers were starting tobarleymeteoritesMODUNC5.APX^ 24317)MM context: After the disastrous 1929 stock market crash, a lot offormerly prosperous financiers saw their future prospects suddenlyturn17)MM target: black^north17)MP target: hopeless solitary17)MS target: illegal ceramic17)MU target: glum^jade17)ML context: Fashion is fickle. At different times, both sophisticatedsociety and rebellious youth have focused their tastes around thecolor17)ML target: black^sound18)MM context: The studio audiences seemed to love the new game show host ,but the critics and a number of people who watched the show feltthat he was too18)MM target: plastic^crowded18)MP target: artificial segregated18)MS target: short level18)MU target: suave^milky18)ML context: One of the biggest problems faced by currentenvironmentally conscious governments is what can be done to stopthe production of throwaway18)ML target: plastic^gardens19)MM context: Alice's playmates thought it was funny to pull thebeetle's legs off one by one and watch it struggle to walk, butAlice thought it was19)MM target: sick^rear19)MP target: horrible^centered19)MS target: inspiring distilled19)MU target: macabre^rumpled19)ML context: It is often a difficult challenge for working couples torearrange their day in order to get adequate home care when theirchild is19)ML target: sick^huge20)MM context: There has been a lot of inefficiency in our administrationof funds. Our new policy is to manage all of our funding under asingle20)MM target: umbrella^traveler20)MP target: agency breath20)MS target: bureaucracy^diffraction20)MU target: federation crawlspace20)ML context: It was a grey, rainy day and Harriet dozed on the busuntil she reached her stop. After she had gotten off, sherealized that she had forgotten her20)ML target: umbrella^revision244LEXICAL DECISION - SENLEXOdie Geiger - June, 1992Name:^Major:Class credit to (instructor)?:^Is English your most familiar language?Please answer the following questions with 'T' if the meaning of the question is the sameas one of the items you were given to read, or "F' if the meaning is different. Some ofthe items in the experiment were presented more than once with different endings; aslong as the question is true of ONE of those you remember, respond 'T'. None of theseare trick questions, for example, names were not switched between items, and incidentshave not been combined.1. ^ Alice's playmates tried to get beetles to fight over bits of grain.2.^ The politician's presentation was smooth, but his facts seemedsuspicious.3.^ The new game show host impressed the critics as highly talented.4.^ It was felt that funds could best be administered under a singleagency.5.^ Tim's teacher's thought that he was manipulative and shallow.6.^ South America is a cyclist's dream.7.^ Edmund's newest novels were masterpieces when compared to hisearlier work.8.^ All of Einstein's ideas were spectacularly successful.9.^ Corrine was hungry after a busy morning of classes.10.^  Jerry's friends found him in a bar, celebrating how well his speechwas received.11.^ The worker's dissatisfaction with management meant that it wastime to call a strike.12.^ Muriel and Howard agreed about nearly everything exceptpolitics.245^13. ^ Bianca saw a great deal of similarity between her life and hermother's life.^14.^ Don was interested in the French Nobel Prize winner's chemistrylecture, because he had been working on the same ideas himself.15.^ Nigel felt that Professor Roger's theories were not well thoughtout.16.^ The lawyers were able to make a clear case against the drugdealers.17.^ Kevin impressed Elizabeth with his knowledge of cars andbaseball.18.^ Steve tried, but failed, to establish a good relationship with hismother-in-law.19.^ Toby gets completely involved in mystery novels by her favouriteauthor.20.^ The committee decided to wait before evaluating Laura'sperformance because she was on a maternity leave.21.^ Jack was nervous before his audition for the producer.22.^ In the United States each immigrant group tends to maintain itsown culture.23.^ Andrea found it frustrating to try to teach students who livesrevolved around personal concerns.24.^ The Department of Education was determined to stand firm ontheir plans for funding cutbacks.25.^ The coach thought he could improve his team's performance byadding some new players.26.^ Margie is a good friend, but she tends to be moody.27.^ Warren knew that it would be impossible to find a simple solutionto all his problems.28. ^ Julie never tired of entertaining a steady stream of visitors andtheir pets.246^29. ^ Tina felt that drama was only worth seeing if it carried somemeaningful message.^30. ^ Pam told Helen that she had planned to elope with Max, but herparents talked them out of it.31. ^ The recycling committee had trouble finding buyers for thenewspapers they collected.32. ^ Carmen made the mistake of getting drunk and acting like theothers instead of thinking for herself.Metaphor Comprehension - 247APPENDIX F248LEXICAL DECISION - DEGRADED CONTEXT - SENLEX9December, 1992This experiment consists of a series of short sentence contexts in which the finalword is missing. Your task will be to choose which of two words completes the intendedmeaning of the sentence, and then to indicate your choice by pressing a buttoncorresponding to the chosen word as quickly as possible.The sentence contexts will be displayed all at once on a computer monitor. Youare to read the sentences, and once you have understood their meaning, press thebottom buttons on the box as indicated. This will remove the context and present youwith two words, side by side, on the video screen. Read them both, and then presseither the right or left button on the top of the box corresponding to the position of theword. Once you have made your response, a new sentence context will automaticallyappear on the screen.As the presentation continues, you will notice that some of the contexts arerepeated. However, the endings for these sentences will be different. It is important tothe success of this experiment that you read through every context each time it appears,even if you recall it from a previous display. There will be a short pencil and paper quizfollowing the experiment that you will find easy to answer as long as each context isunderstood in terms of its endings.If you are ever unfamiliar with the meaning of one or both of the wordspresented, or if you are unable to decide between the two, or if neither of the wordsseems to make sense to you in terms of the context, just make a guess and continue onwith the experiment. You should try to work at a steady pace, and to be as accurate aspossible in making your choice.The entire experiment consists of 200 trials of sentence contexts followed by thequiz, and should take about one hour to complete. You will receive one hour of classcredit for your participation. Thank you for your interest.Odie Geiger,Research AssistantDr. Lawrence Ward,Supervising Professor249DEADUNC6.APXEXPERIMENT 6 - DEAD METAPHORSNote: In this appendix, correct targets are listed first, followed bymatched incorrect targets. In the actual experiment, the correct stimuluwas randomly selected to appear on either the right or left side of thescreen.context: Finding a buyer was an unexpectedtarget: bottleneck^paraphrasetarget: difficulty atmospheretarget: expense^welcometarget: impediment hypodermiccontext: They were made so cream would collect in thetarget: bottleneck^sketchbookcontext: He was personable and verytarget: bright^battletarget: smart validtarget: rich^gaintarget: astute maltedcontext: The sky was clear andtarget: bright weightcontext: His mother-in-law remainedtarget: cold^movetarget: indifferent^seventeenthtarget: shy^partarget: aloof plaidcontext: She knew Norway would betarget: cold^sortcontext: They had opposingtarget: positions^electionstarget: opinions suppliestarget: loyalties footnotestarget: stances^oatmealcontext: She felt competent when she assumed the steps andtarget: positions employeescontext: All he talked about was cars. She had never met anyone sotarget: shallow^pasturetarget: superficial^microscopictarget: masculine scratchedtarget: insipid^twirledcontext: The water is warm andtarget: shallow^amended1) DD1)DD1) DP1) DS1)DU1)DL1)DL2)DD2) DD2) DP2)DS2)DU2) DL2)DL3)DD3)DD3)DP3)DS3)DU3)DL3)DL4)DD4)DD4)DP4)DS4) DU4)DL4)DL5)DD5)DD5)DP5)DS5)DU5)DL5)DL6)DD context: He spent6)DD target: formula6)DP target: solutionyears searching for a simpletensionrailroad6)DS target: faith^staff6)DU target: tenet snout6)DL context: Her discovery removed stains. They offered her money for her6)DL target: formula^housingDEADUNC6.APX^ Friday, June 4, 1993^ 2507) DD7) DD7) DP7)DS7) DU7)DL7) DLcontext: I can turntarget: supporttarget: comforttarget: moneytarget: sustenancecontext: They maketarget: supportto him forlookingregionsdeathcasualtiessure superstructures have adequatepicture8)DD context: They maintained their plans were8)DD target: firm^cars8)DP target: definite^swimming8)DS target: progressive respiratory8)DU target: unequivocal^chauffeured8)DL context: They have textures ranging from soft to8)DL target: firm^pool9)DD context: She hated entertainment that was too9)DD target: deep^unit9)DP target: profound^released9)DS target: expensive separated9)DU target: abstruse^deflated9)DL context: The well was almost 50 meters9)DL target: deep^jazz10)DD context: He found the lecture difficult to10)DD target: grasp^tumor10)DP target: understand^population10)DS target: translate dimension10)DU target: comprehend^accelerate10)DL context: Her installed handrails so she would have something to10)DL target: grasp^quote11)DD11) DD11)DP11)DS11) DU11) DL11)DL12)DD12)DD12)DP12)DS12)DU12)DL12)DLcontext: She finds novels sotarget: absorbing^shufflingtarget: interesting responsibletarget: difficult^committeetarget: engrossing unsinkablecontext: They are developing a material that is highly oiltarget: absorbing^threadingcontext: They are terrified of riding where roads have inadequatetarget: shoulders^decisionstarget: edges^kneestarget: restaurants^metaphysicstarget: verges rivalscontext: She never went out without a shawl over hertarget: shoulders^elections251DEADUNC6.APX13)DD context: When Pam eloped, Helen was13)DD target: shocked^blowing13)DP target: surprised organized13)DS target: happy13)DU target: appalled13)DL context: If you use13)DL target: shockedglassprolifican appliance in the bathtub you will getbroader14)DD14)DD14)DP14)DS14)DU14)DL14)DLcontext: They had nevertarget: depressedtarget: unhappytarget: drunktarget: disheartenedcontext: In the summertarget: depressedseen him sorenderingstainedmagictransluscentthe water level becomes dangerouslyrespected15)DD context: He worried because the smell was so15)DD target: strong^beyond15)DP target: intense concert15)DS target: accepted opposite15)DU target: pungent^tubular15)DL context: He exercised to keep himself15)DL target: strong inside16)DD context: Her life16)DD target: parallel16)DP target: similar16)DS target: cursed16)DU target: akin16)DL context: It must16)DL target: paralleland her mother's wereregionalearlierporoussurfhave been difficult to keepenormousthe tracks17)DD context: The facts were17)DD target: clear^north17)DP target: obvious musical17)DS target: unusual highest17)DU target: indisputable^aeronautical17)DL context: The water is warm and17)DL target: clear^alone18)DD18)DD18)DP18)DS18)DU18)DL18)DLcontext: Hertarget: flytarget: runtarget: itchtarget: jogcontext: Thetarget: flyfeet began topatcutslamribrobin learned tobidDEADUNC6.APX^ 25219)DD context: His reasoning seemed so19)DD target: weak^hurt19)DP target: vague brave19)DS target: silly lined19)DU target: imprecise^sleepless19)DL context: The hostages were extremely19)DL target: weak^bare20)DD context: When he married her, she had no idea he was so20)DD target: calculating^fragmentary20)DP target: shrewd^barred20)DS target: poor grey20)DU target: wily^dank20)DL context: His balances came out wrong after hours of20)DL target: calculating^adolescentsMODUNC6.APX^ 253EXPERIMENT 6 - METAPHORSNote: In this appendix, correct targets are listed first, followed bymatched incorrect targets. In the actual experiment, the correct stimuluwas randomly selected to appear on either the right or left side of thescreen.context: They decided to wait because she was stilltarget: green^seventarget: inexperienced^environmentaltarget: recovering synonomoustarget: unversed^embossedcontext: She thought he would notice but to him it was stilltarget: green teethcontext: Making an organization run requires elaboratetarget: choreography^astrophysicstarget: planning^religiontarget: rules youthtarget: designing figurinescontext: They were thrilled with such innovativetarget: choreography^bewildermentcontext: By noon, she neededtarget: fueltarget: foodtarget: exercisetarget: sustenancecontext: He kept atarget: fuel1)MM1)MM1)MP1) MS1)MU1) ML1)ML2)MM2)MM2)MP2)MS2)MU2)ML2)ML3)MM3)MM3)MP3)MS3)MU3)ML3)MLsumshairmachineshighlightsrecord of his car's consumption ofcatsto do well, but he wasclawedbiggestgarlandedvexatious4)MM context: He wanted4)MM target: wobbly4)MP target: nervous4)MS target: hiccuping4)MU target: flustered4)ML context: Her ladder was4)ML target: wobbly^pitied5)MM context: In the U.S., cultures have become5)MM target: interwoven^morticians5)MP target: combined hydrogen5)MS target: rebellious bequeathed5)MU target: conjoined^subtitled5)ML context: Shebought fa bric in which strands had been5)ML target: interwoven^reconvened6)MM context: Not even6)MM target: gems^6)MP target: valuable6)MS target: original6)MU target: laudable6)ML context: They can6)ML target: gemsEinstein's ideas were allpegschangingreturnedbuttonedbuy artificial materials, believing they arehubsMODUNC6.APX^ 2547)MM7)MM7)MP7)MS7)MU7)ML7)MLcontext: She told himtarget: juggletarget: managetarget: concealtarget: superintendcontext: He found thetarget: juggleit was too much todilutebundlebreathefirecrackerdiversion he needed when he learned tostaple8)MM context: The project8)MM target: bear8)MP target: burden8)MS target: farce8)MU target: tribulation8)ML context: They wanted8)ML target: bearwas becoming avetolawyerleashrequisitionto photograph arust9)MM context: Her nature9)MM target: sunny9)MP target: cheerful9)MS target: simple9)MU target: genial9)ML context: The skies9)ML target: sunnyis always gentle andbladedeceasedhigherdottedbecamealoud10)MM10)MM10) MP10)MS10)MU10)ML10)ML11)MM11)MM11)MP11) MS11)MU11)ML11)ML12)MM12)MM12)MS12)MU12)MP12)ML12)MLcontext: She suspected the data had beentarget: cooked^caringtarget: changed talkingtarget: stolen cosmictarget: contrived^nocturnalcontext: She didn't stop until they were peeled andtarget: cooked rearedcontext: His teamtarget: bloodtarget: playerstarget: uniformstarget: athletescontext: He had totarget: bloodneeded some newtreessamplesfeathersexcerptsbe insensitive to thetruthsight ofcontext: Each night she went home feelingtarget: empty^thicktarget: sweaty woodedtarget: ineffectual^aerodynamictarget: useless^flowingcontext: He hurried to the auditorium but found ittarget: empty idealMODUNC6.APX^ 25513)MM context: He and his dog were arriving. She couldn't stand another13)MM target: assault^profile13)MP target: guest stick13)MS target: animal^career13)MU target: caller pillar13)ML context: They were left to defend against an enemy13)ML target: assault^collage14)MM context: She wondered how she had been able to follow the14)MM target: wave^salt14)MP target: crowd grass14)MS target: tune^doll14)MU target: horde blush14)ML context: They worry about a tidal14)ML target: wave^bond15)MM context: They complained his work had become15)MM target: stagnant^tapestry15)MP target: dull^sand15)MS target: childish^tropical15)MU target: hackneyed laminated15)ML context: The pond where he swam was now15)ML target: stagnant^perverse16)MM context: It was time to strike. The workers were making16)MM target: noises^barley16)MP target: complaints^meteorites16)MS target: mistakes arteries16)MU target: innuendoes^tombstones16)ML context: They tried to describe their car's odd16)ML targets: noises^storms17)MM context: They saw their prospects turn17)MM target: black^north17)MP target: hopeless solitary17)MS target: illegal^ceramic17)MU target: glum jade17)ML context: Both have focused their tastes around17)ML target: black^sound18)MM context: Critics felt he was too18)MM target: plastic^crowded18)MP target: artificial segregated18)MS target: short level18)MU target: suave^milky18)ML context: The problem is to stop the production of throwaway18)ML target: plastic^gardensMODUNC6.APX^ 25619)MM context: They pulled its legs off. She thought it was19)MM target: sick^rear19)MP target: horrible^centered19)MS target: inspiring distilled19)MU target: macabre rumpled19)ML context: It is difficult to get care when a child is19)ML target: sick^huge20)MM context: Our policy is to manage our funding under a single20)MM target: umbrella^traveler20)MP target: agency breath20)MS target: bureaucracy diffraction20)MU target: federation^crawlspace20)ML context: She had forgotten her20)ML target: umbrella revision257LEXICAL DECISION - DEGRADED CONTEXT - SENLEX9December, 1992Name: ^Major:Class credit to (instructor)?:^Is English your most familiar language?:^Please answer the following questions with "T" if the meaning ofthe question is the same as one of the items you were given toread, or "F" if the meaning is different. Some of the items in theexperiment were presented more than once with different endings; aslong as the question is true of ONE of those you remember, respond"T".1. She watched it walk briskly away.2. She was suspicious of some of the data.3. He impressed the critics with his talent.4. It was felt that funds could best be administeredthrough a single agency.5. He was appeared to be personable, but he wasmanipulative.6. They enjoy riding on narrow, busy roads.7.^ They complained his work was too minimalist andliterary.8.^All of Einstein's ideas were spectacularly successful.9.^ She was hungry after a busy morning.10.^His friends joined him in celebrating his success.^11.^ The worker's frustrations made it necessary for them tocall a strike.12.^Each of them had opposing views.13.^ She saw a great deal of similarity between her life andher mother's.14.^ He found the lecture extremely interesting andinformative.15.^ He thought the theories were not well thought out.16.^ Thee facts of the case were straight-forward.17.^ She was impressed by his knowledge of cars.258^18.^ He tried, but failed, to establish a good relationshipwith his mother-in-law.^19.^ She gets completely involved in the novels she reads.20.^ They decided to wait before evaluating her performancebecause she was on maternity leave.21.^He wanted to do well, but he was nervous.22.^ In the U.S., cultures have tried to maintain their ownidentity.23.^ She went home frustrated and exhausted each night.24.^ They were determined to stand firm on their originalplans.25.^ The team would benefit by getting some new players.26.^ She is usually moody, and often mean.27.^ He found the answer to all his problems in a fortunecookie.28.^ She never tired of entertaining her friends and theirpets.29.^ She felt drama was only worth seeing if it carried ameaningful message.30.^ When Pam wanted to elope, Helen talked her out of it.31.^ They had trouble finding a buyer.32.^ She wondered how she could have gone along with theothers, instead of thinking for herself.

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0098794/manifest

Comment

Related Items