Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Young children’s comprehension of comparative adjectives: problems of experimental design and interpretation Barty, Naomi 1976

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

Item Metadata

Download

Media
831-UBC_1976_A6_7 B38.pdf [ 3.8MB ]
Metadata
JSON: 831-1.0100152.json
JSON-LD: 831-1.0100152-ld.json
RDF/XML (Pretty): 831-1.0100152-rdf.xml
RDF/JSON: 831-1.0100152-rdf.json
Turtle: 831-1.0100152-turtle.txt
N-Triples: 831-1.0100152-rdf-ntriples.txt
Original Record: 831-1.0100152-source.json
Full Text
831-1.0100152-fulltext.txt
Citation
831-1.0100152.ris

Full Text

YOUNG CHILDREN'S COMPREHENSION OF COMPARATIVE ADJECTIVES: PROBLEMS OF EXPERIMENTAL DESIGN AND INTERPRETATION by NAOMI BARTY B.A. (Hons.), University of British Columbia, 1972 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE FACULTY OF GRADUATE STUDIES Department of Paediatrics Division of Audiology and Speech Sciences We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA August, 1976 (c\ Naomi Barty, 1976 THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in In presenting th i s thes is in pa r t i a l fu l f i lment of the requirements for an advanced degree at the Univers i ty of B r i t i s h Columbia, I agree that the L ibrary shal l make it f ree ly ava i l ab le for reference and study. I fur ther agree that permission for extensive copying of th is thesis for scho lar ly purposes may be granted by the Head of my Department or by his representat ives. It is understood that copying or pub l i ca t ion of th is thes is for f inanc ia l gain sha l l not be allowed without my writ ten permission. Department of rpctSJJJL^LA^ The Univers i ty of B r i t i s h Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Date Au 6 ABSTRACT Previous investigations have shown that children between the ages of three and five years give consistently correct responses to re-quests for one of a pair of quantities which has more elements and in-correct responses to requests for the one containing less. On such a basis i t has been hypothesized that the word more is properly understood before the word less and, at some point in the acquisition of these terms, less is understood to have the meaning of more. However, such a response bias is not demonstrable in a l l of the studies which have required children to deal with the terms more and less. As a consequence, the various experimental methodologies are examined for differences which may have accounted for the disparity of results, i.e., for the presence or absence of the response bias described. The present study investigates the performance of thirty children between the ages of 2; 6 and 4;6 on comparisons which involve the terms more/less and longer/shorter. Two types of comparison were made: (1) the choice of one stimulus in response to the words more and less and longer and shorter, and (2) the choice of one of the terms in the comparison of one stimulus with a given standard. Very few of the children tested demonstrated a response bias in favour of the positive terms, more and longer, and their scores on the comprehension of more and less were found to be highly dependent on the design of the task. The 'Semantic Feature Hypothesis' proposed by E. Clark (1973) suggests that the confusion of polar opposites (i.e., opposing comparative i i adjectives along the same dimension) is part of a more general principle of semantic development and that a l l children go through a stage in which the same meaning is attributed to more and less, longer and shorter. The present findings, on the other hand, reveal that almost a l l the children tested give either consistently correct responses to both terms or incon-sistent responses, that children rarely respond consistently incorrectly to either member of the pairs, and that when they do, their performance on a production task indicates either comprehension of both words or non-comprehension of both, but does not support the suggestion that the two words are treated as synonyms. The inadequacy of the tasks used as a measure of comprehension of comparative adjectives constitutes a major di f f i c u l t y for the 'Semantic Feature Hypothesis', insofar as its formula-tion was dependent on the assumption that the performance of children in such situations was a valid indication of their understanding of the words in question. The results of the present investigation suggest that such a claim is untenable and that careful control of experimental design is particularly important in the adequate assessment of cognitive-linguistic capabilities. i i i TABLE OF CONTENTS ABSTRACT i i TABLE OF CONTENTS iv LIST OF TABLES v i i LIST OF FIGURES v i i i ACKNOWLEDGMENTS . ix CHAPTER 1 INTRODUCTION 1 1.0 Introduction 1 1.1 Review of the Literature 3 1.11 The Acquisition of more and less . . . . 3 1.12 Cognitive Correlates of the Acquisition of Comparative Terms 11 1.13 Bases for the Apparent Confusion of more and less 17 1.14 The Experimental Assessment of Comprehension of more and less: A Comparison of Methodologies 23 1.2 Summary and Statement of the Problem . . . . 32 2 EXPERIMENT I 36 2.1 Method . 36 2.11 Subjects 36 2.12 Materials 36 2.13 Procedure 38 iv Page CHAPTER 2 2.14 Scoring 40 2.2 Results 40 2.21 More vsless; longer vs shorter . . . . 40 2.22 Age Effects 41 2.23 Stimulus Effects 43 2.24 Summary 43 3 EXPERIMENT II 46 3.0 Introduction . . . . 46 3.1 Method 46 3.11 Subjects . . . . 46 3.12 Materials 46 3.13 Procedure 47 3.14 Scoring 49 3.2 Results 49 3.21 More vs less; longer vs shorter . . . . 49 3.22 Age Effects 50 3.23 Stimulus Effects 50 3.3 Combined Results of Experiments I and II . . . 54 3.30 Introduction 54 3.31 Age Effects 54 3.32 Stimulus Effects 56 3.33 Linear Ordering . 59 v Page CHAPTER 4 DISCUSSION 60 4.0 Introduction 60 4.1 Discussion of Results 60 4.11 Experiment I 60 4.12 Experiment I I 61 4.13 Combined Results and Comparison of Experiments I and I I 63 4.2 Comparison of the Present Results with those of Previous Investigations 64 4.3 Summary and Conclusions 70 REFERENCES . . . • 72 v i LIST OF TABLES Page TABLE 1 Summary of Previous Research by Task Classification . . 28 2 Experiment I. Mean scores and percent correct for more/less and longer/shorter tasks 42 3 Experiment I. Correct responses (7») on Tasks M/L-l, M/L-2 and L/S-l for four age groups 42 4 Experiment I. Number of and percentage correct responses for each stimulus on Tasks M/L-l and M/L-2 . . . . 44 5 Experiment II. Mean scores and percent correct for Type A and B stimuli for more/less tasks and for longer/shorter tasks 51 6 Experiment II. Correct responses (7o) on Tasks M/L-l, M/L-2 and L/S-l for four age groups 52 7 Experiment II. Number of and percentage correct responses for each stimulus on Task M/L-l and Task M/L-2 53 8 Experiments I and II. Correct responses (%) on Tasks M/L-l,. M/L-2 and L/S-l for four age groups . . . . 55 9 Experiments I and II. Correct responses (%) to Type A stimuli on Tasks M/L-l and M/L-2 57 10 Experiments I and II. Correct productions (7») of more for three stimulus pairs in Task M/L-2 for four age groups 57 11 Experiments I and II. Mean correct scores on Task M/L-l for thirty subjects, categorized according to behavior on linear ordering 58 V I I LIST OF FIGURES Page FIGURE 1 Stimuli for more/less Tasks in Experiment I (Pairs 1-3) and for more/less Type A Tasks in Experiment II (Pairs 1-4) • • 37 2 Stimuli for more/less Type B Tasks in Experiment II . 48 v i i i ACKNOWLEDGMENTS I would like to thank a l l those who directly or indirectly contributed to the completion of this thesis. In particular, I wish to express my thanks to Dr. John Delack for his encouragement and his conscientious supervision of the project, and Dr. Bob Frender for his advice and many helpful suggestions. I would also like to thank Wendy Goodbrand and Jean Ronalds for their assistance in finding suitable subjects. Thanks are also due, of course, to the children who served as subjects and their families for their cooperation. Finally, I would like to thank Katharine Cook for her careful typing of the final draft. ix CHAPTER 1 INTRODUCTION 1.0 Introduction Recent interest in the relationship between cognitive development and the acquisition of language has focused primarily on the cognitive bases for advances in semantic and syntactic development. This is ultimately the most valuable application of knowledge of the development of cognition for those interested in the development of language, but there is also need at present to examine the contribution of cognitive differences to the way in which a child's treatment of some tasks differs from that which would be expected of an adult. Insight into the acquisition of a f i r s t language is often sought by means of manipulative experiments which attempt to set up environmen-tal conditions such that a particular response is contingent upon know-ledge of a particular concept or principle. It is legitimate and often necessary to create a r t i f i c i a l conditions in order to examine the l i n -guistic competence of children, but in interpreting their behavior in situations created to assess a particular linguistic a b i l i t y , the danger of generating false impressions through failure to take into account the cognitive differences between the language-learning child and the adult must be recognized and avoided. A case in point is the controversy which began in 1968, when Donaldson and Balfour discovered, in an experiment designed to assess 2 young children's comprehension of the words more and less, that the majority of their subjects responded to the two words as i f they believed them to be synonymous. Since that time, research and theorization have been based on the assumption that the task used was a definitive assess-ment of children's comprehension of more and less. These data have been used in support of a theory which holds that words such as more and less are treated as synonyms during a stage in the acquisition of meaning when the lexical entries for the words are not complete (cf. E. Clark, 1973). Since this theory is intended to represent a general principle of semantic development, i t is crucial to determine whether the responses given by children to the type of tasks generally presented are indeed indicative of their comprehension of the words under investigation. Children's comprehension of the comparative adjectives more and less has been selected for examination here, since research relating to these words has produced conflicting and often surprising results, for which both linguists and psychologists s t i l l seek c l a r i f i c a t i o n . t 3 1.1 Review of the Literature 1.11 The Acquisition of more and less The tendency for young children to respond to comparative sen-tences containing the word less in the same way as those containing the word more (Donaldson and Balfour, 1968) has led to speculation regarding children's understanding of these and other comparative terms. One approach toward an explanation of these data has involved a semantic analysis of the words. H. Clark (1970) points out that more has the meaning of 'an additional quantity o f , as well as the comparative mean-ing, as in 'more than another'. In the sense of 'more o f , more is among the f i r s t words used by children in the one- and two-word stages of language acquisition (Bloom, 1970). H. Clark (1970) hypothesizes that children f i r s t learn the word in neither a comparative nor an additive sense. Thus; the sentence, 'Do you want more food?' is interpreted as 'Do you want some food?'. The comparative interpretation of the sentence (i.e., 'Do you want more food than you have been given so far?') purpor-tedly depends on the knowledge of two underlying base strings, viz., 'You want some food' and 'You have been given some food so far', where the second string is pragmatically implicit. The child, who has not discovered the implicit comparison, ostensibly understands the sentence to have only the meaning of the f i r s t , explicit string. The derivational structure of English comparatives i s , however, not well established, and a case has been made for the derivation of comparative sentences from a single base string (Campbell and Wales, 1969). Furthermore, Bloom's (1970) analysis of children's one- and two-4 word utterances indicates that more is used to signify recurrence of an object or event, or another instance of an object. If more were under-stood to mean 'some1, this would be reflected in the use of the term to request or comment on more X in the absence of an original occurrence of X, which rarely appears in the utterances recorded by Bloom. Weiner (1974) investigated the relationship between the additive and comparative meanings of more. She hypothesized that i f more were understood only in the sense of recurrence, the observed addition of elements to one of two sets would increase the probability of that set's subsequently being chosen as the one with more. There was, however, no significant difference in the number of correct responses between static inequality and transformational conditions. E. Clark (1973) interprets Donaldson and Balfour's (1968) data as an indication that children treat relational opposites, such as more/  less, as synonyms before they learn to contrast them. The children in Donaldson and Balfour's study gave a significantly greater number of correct responses to questions containing the word more than to the same questions containing the word less. Further investigation (Donaldson and Wales, 1970) showed a similar asymmetry in favor of the positive member of pairs of dimensional terms, such as taller/shorter and bigger/  smaller. In light of her interpretation of the data of others, Clark (1973) formulated the 'Semantic Feature Hypothesis', according to which children acquire the semantic features of words one at a time, in a predictable hierarchical pattern, from the most general to the most specific features. 5 She concludes that in learning relational pairs, children acquire features that are common to the two words f i r s t . In the acquisition of more and less, for example, the feature amount is f i r s t attached to both words with a positive coefficient. At this stage the words are used synony-mously. The child then adds [+polar] to the entry for more, and assumes that less also contains this feature. At this stage, both words have the meaning of more. The words are used contrastively when the child learns that 'less refers to the other end of the scale' (1973, p. 91) and attaches the feature [-polar] to that term. The +/- notation is conven-tionally used to indicate the presence/absence of an attribute or feature; in this case, [-polar] would mean 'not polar 1. Clark, however, refers to the negative member of a pair of dimensional terms, to which the feature [-polar] is attached, as the 'negative-pole term'. It is assumed, therefore, that [-polar] is meant to mean 'negative pole', rather than Jnot polar', even though Clark otherwise uses the +/- nota-tion in the conventional manner. If confusion with sentences containing comparative adjectives is a consequence of incomplete or inaccurate lexical entries for the words, then i t might be expected that this confusion would be apparent whenever children deal with these terms. The same response patterns have not always been observed, however, when tasks which require comprehension of more and less and other comparative terms have been administered (Griffiths, Shantz and Sigel, 1967; Inhelder, Sinclair and Bovet, 1974; Weiner, 1974; Eilers, Oiler and Ellington, 1974). Needless to say, a search for consistent factors in situations in which children can and cannot differentiate instructions containing contrasting relational terms 6 should precede hypotheses about the bases of their d i f f i c u l t i e s . Donaldson and Balfour (1968) presented pairs of stimuli which showed quantitative equality and inequality. The subjects, 3.5- to 4-year old children, were required to choose between two cardboard trees on which cardboard apples were hung. The questions asked were: 'Which tree has more apples?' and 'Which tree has less apples?'. The subjects were also required to create inequalities in response to the instruction to 'put more (less) apples on one tree than the other', and then to identify the one with more and the one with less. In each of these situations children consistently chose the greater quantity in response both to requests for the one with more and the one with less. As a result, the number of correct responses to questions containing the word more was significantly greater than the number of correct responses to the corresponding ques-tions containing the word less. Palermo (1973) attempted to replicate Donaldson and Balfour's (1968) research using the same stimuli and also investigated the compre-hension of more and less with continuous quantities. Children between the ages of 3 and 5 years were given the same tasks as those administered by Donaldson and Balfour, as well as additional, similar tasks with glasses of water in place of trees and apples. The questions for this task were of the same format as those for the f i r s t task: 'Which glass has more (less) water?'. The results for the two tasks were similar to each other and to those of Donaldson and Balfour: Children consistently chose the greater quantity in response to requests for both more and less, a tendency which resulted in a significantly greater number of correct 7 responses to more than to less questions. On this basis, Palermo hypothe-sized that i f children 'did not know the concept [of less], they treated i t as i f i t were a synonym of more' (Palermo, 1973, p. 214), although there was no empirical support for such a conclusion. Palermo (1973) also studied the performance of kindergarten, f i r s t and second grade children on the same tasks. Although the ages of these children are not given, second grade children in American public schools are generally between six and eight years of age. Some of the children, even in this oldest group, s t i l l had d i f f i c u l t y with the word less. G r i f f i t h s , Shantz and Sigel (1967) presented 4- to 5-year-old children with one row of lollipops as a standard for comparison, followed by a second row which contained more (or less) lollipops than the stan-dard. An attempt was f i r s t made to have children use more and less spontaneously with the question: 'What can you t e l l me about these two sets of lollipops?' If relational terms were not used in the response, the child was asked: 'Does this set of lollipops have more lollipops, less lollipops, or the same number of lollipops as this set?' The terms more and less were significantly more often e l i c i t e d than spontaneously emitted by the subjects. Comparative judgements using the words more and less were 727,, and 80% correct, respectively. Judgements of relative length, under the same conditions were 100% correct for longer and 96% correct for shorter. Griffiths, Shantz and Sigel's task was essentially a production task, while those of Donaldson and Balfour (1968) and Palermo (1972) dealt with comprehension. In the latter studies, neither stimulus was presented 8 as a standard, while in the former case, children were given one stimulus as a standard and were asked to specify whether the difference between i t and the item of comparison was in a positive or negative direction. Inhelder, Sinclair and Bovet (1974) gave 4.5- to 8-year-old children instructions such as: 'Give more modelling clay (pate) to the g i r l and less to the boy', and 'Give the g i r l more marbles than the boy', or 'Give more marbles to the g i r l and less to the boy 1. These were carried out without error even by the youngest children. This exercise, like those of Donaldson and Balfour (1968) and Palermo (1973), is basically a test of comprehension, but i t differs from their tasks in that both the positive and the negative poles are specified in the instructions. The children were required to create inequalities, as they were for one of Donaldson and Balfour's tasks. However, both terms were used in the instructions, and the distribution of inequalities was specified by the experimenter, while Donaldson and Balfour asked subjects to identify more and less after distributing the items. Weiner (1974) found that 3-year-old children made more correct choices between two rows (or 'parts') of toys in response to the ques-tion, 'Which part has more ?', than to the corresponding question containing the word less. Children performed at chance level in response to less questions. Two-year-old children in this study performed close to chance levels for questions containing both more and less, except under one condition in which one row contained the maximum number of toys (i.e., the space was f i l l e d ) . Weiner hypothesizes that children may have an intermediate interpretation of more as 'many1. 9 Donaldson and McGarrigle (1974) demonstrated the use of 'fullness' as a criterion for applying the term more to a quantity. They found that the responses of about one-third (14 out of 40) of their subjects between the ages of 3;0 and 5;2 to the question: 'Are there more cars on this shelf or more cars on this shelf?' depended on whether or not the cars occupied a l l the spaces in the garages that contained them. A comparison between a row of four and a row of five cars was made under two condi-tions: (1) four cars contained in a four-space garage and five cars contained in a six-space garage, and (2) the same rows of cars with the garages removed. A l l children chose the row of five cars as more in condition 2, but in condition 1, when the row of five was contained in a six-space garage, one-third of the children chose the row of four cars in a four-space garage as more. Donaldson and McGarrigle conclude that when possible some children use 'fullness' rather than length as a criterion for application of the word more. Palermo (1974) investigated the possible contribution of the types of quantities compared in more/less comprehension tasks to the differences in results obtained by studies which involved comparisons of weight, linear arrays, nonlinear arrays, and continuous quantities. Children between the ages of 3.2 and 4.3 years made the following compari-sons: (a) quantities of apples in nonlinear array, (b) quantities of water in glasses, (c) blocks of equal size but different weights, and (d) poker chips arranged linearly and in one-to-one correspondence. For each task the child was presented with two unequal quantities and was required to choose the member of the pair which had less. The number of incorrect choices made did not vary significantly among tasks. Palermo 10 found a bimodal distribution of total scores, with only 3 of the 32 children obtaining scores between 257. and 757., correct. He concludes that this reflects a division between children who knew the concept of less and those who did not. Differences between his results and those of Griffiths, Shantz and Sigel (1967) are attributed to the possible f a c i l i t a t i o n provided by greater syntactic context given the child in the latter study. Palermo does not cl a r i f y , however, how a child who understands less to mean more could be helped by syntactic contexts which are identical for both terms. Reports of asymmetrical scores on comprehension tasks involving dimensional adjectives do not consistently show higher scores for the positive members of pairs of polar opposites. Eilers, Oiler and E l l i n g -ton (1974) found a response bias in subjects between the ages of 2;6 and 3;6 which favored the 'negative' (or 'marked') member of the pairs big/  l i t t l e , long/short and wide/narrow. Further experimentation revealed that this bias existed when the instructions were simply: 'Here are two X's. Give me one.'. It is possible, then, that at least in this case the preference for one member of a pair is the cause rather than the product of the apparent asymmetry in children's a b i l i t y to handle instruc-tions containing positive and negative polar adjectives. In an effort to resolve the apparent conflict between their results and those predicted by the 'Semantic Feature Hypothesis' (Clark, 1973), Eilers et a l . postulate two strategies for dealing with these comparatives: Those who follow strategy A attribute the meaning of the marked member to the unmarked member of the pair; and those who follow 11 strategy B attribute the meaning of the unmarked to the marked member. They hypothesize further that at some stage the child makes a choice between the marked and unmarked mea.ning, thereafter attributing that meaning to both words. They do not suggest why i t would be necessary for a child, for whom both meanings were available, to choose only one of them. The appearance of a response bias in this and other studies does suggest that children develop preferences for either the positive or the negative member of a pair, but there is insufficient evidence to suggest that this preference has a semantic basis. Holland and Palermo (1975) found that 28 of 31 children between the ages of 4;10 and 5;11 who were not able to identify a quantity with less were able to do so with training. The authors report that the appropriate reference for less was learned quickly and easily by these children. They suggest that the less-more confusion may be artifactual -a result of the child's perception of the greater quantity as the better choice regardless of instructions. 1.12 Cognitive Correlates of the Acquisition of Comparative Terms Children's comprehension of the terms more and less is of sig-nificance in the interpretation of Piagetian conservation studies. Concern has been expressed over the contribution of semantic confusion to incorrect responses given by children in conservation tasks (Griffiths, Shantz and Sigel, 1967). If, as theorists such as E. Clark (1973) pro-pose, children understand less at some stage in development to be iden-t i c a l in meaning to more or, as H. Clark (1970) suggests, they interpret more as 'some', then their use of these words would not be a clear indi-12 cation of their understanding of relative quantity. An i l l u s t r a t i o n of the relationship of the meaning assigned to the word more to the develop-ment of the notion of quantity is provided by the following series of experiments. Mehler and Bever (1967) found that 2-year-old children gave correct responses to questions requiring the choice of the member of a pair of rows which contained more objects, and that these children obtained higher scores on the task than did older children (up to the age of 4;11). The children were required to choose the more numerous of two rows of objects which were not in visual one-to-one correspondence. The decrease in scores with age was interpreted as a decrease in the a b i l i t y to conserve quantity - a temporary phase which occurs as a result of the child's learning to associate longer arrays with more components. The 're-acquisition' of conservation, according to Mehler and Bever, depends on the child's learning to ignore perceptual expectancies in cases in which they conflict with logical operations. Piaget (1968), however, interprets Mehler and Bever's results as an indication of the tendency of the children to 'center' perceptually on the density of the row in which the objects are closer together. He also suggests that, since the children observed the experimenter adding elements to one of the rows, the word more may have been interpreted in some cases in the additive, rather than the comparative sense. In a series of experiments with subjects between the ages of 2;3 and 3;10, Piaget obtained results which do not support Mehler and Bever's conclu-sions regarding the a b i l i t y of 2-year-old children to conserve. A l l 13 children in his study chose one of a pair of rows containing equal numbers of elements as more when one row was spread out or crowded together. There was a difference between age groups in the tendency to use the criterion of length in making judgments of quantity. Some of the sub-jects up to the age of 3;2 used crowding and length alternately as c r i t e r i a , whereas a l l but one of the older children judged on the basis of length only. Piaget interprets these results in terms of the development of spatial concepts. Before length is considered, children's conception of space is limited to topological structures based on proximity, separa-tion, enclosure and frontiers. Thus crowding, or the proximity of elements, is a more primitive criterion than length for the evaluation of quantity. Piaget points out that the consistency of answers to Mehler and Bever's tasks is affected by the ratio of the lengths of the two rows. When children were presented with a situation in which a row of four elements was 1.6 times longer than a row of six elements, a l l children between 2;4 and 2;7 correctly chose the shorter, denser row as the one with more, whereas a row of six elements which was five times longer than a row of four elements was correctly chosen by only 69 percent of sub-jects between 2;4 and 2;9 as the one with more (Mehler and Bever, 1967). Piaget hypothesizes that subjects at an intermediate stage between evaluation by crowding and evaluation by length would be more inclined to be struck by length in the second situation than the f i r s t , and would choose the shorter, denser row in the former and the longer, less dense row in the latter situation. Piaget also draws attention to Mehler and Bever's finding that even when one row was five times longer than the 14 other and contained more elements, some two-year-old children s t i l l chose the shorter, denser row as having more. Bever, Mehler and Epstein (1968) argue that children understand more as a relative, rather than an absolute term on the basis of their consistently correct choices of the row with more elements even when they did not observe the experimenter adding to and crowding together one of the originally equal rows. Although this supports the argument that children were not responding to the addition of elements, i t weakens their contention that these children demonstrate the a b i l i t y to conserve quantity; as Beilin (1968) notes, i f the transformations are not observed, and i f the elements are not in visual one-to-one correspondence, children who cannot count must be responding to perceptual cues of either length or density. Piaget (1968) maintains that the terms more and less do not have a consistent meaning for very young children and that answers given by 2-year-old children in his study varied when the same questions were asked more than once. He found that absolute terms, such as a lot and a l i t t l e produce more consistent responses than more and less. As Mehler, Bever and Epstein (1968, p. 293) suggest, 'linguistic differences are not explanations of cognitive differences, but reflec-tions of them'. Thus, the child's understanding of the word more would be, in some sense, a reflection of the notion of quantity and of the a b i l i t y to establish relations between stimuli. The experiments per-formed by Mehler and Bever (1967), Bever, Mehler and Epstein (1968) and Piaget (1968) demonstrate that between the ages of two and four years 15 changes occur in children's methods of judging quantities and of making comparisons. Piaget's suggestion that children might use the word more in an additive rather than a comparative sense was not supported by Bever, Mehler and Epstein's (1968) comparisons of the number of correct responses with and without the observed addition, nor by Weiner's (1974) investigation of the effect of adding and subtracting elements on the understanding of more and less. It is evident from the apparent varia-tion in the c r i t e r i a used for assignment of the label more, as demon-strated both by Piaget (1968) and by Donaldson and McGarrigle (1974), that understanding of quantity is not operational, but is based on static perceptual cues, such as proximity of elements, lengths of rows or fullness of space. Sinclair (1969) linked the tendency to use comparative, rather than absolute, terms with the development of conservation in children. She found that when asked to describe unequal quantities of marbles, a l l children who demonstrated the a b i l i t y to conserve used comparative terms, whereas ninety percent of the children who could not conserve used abso-lute terms (e.g., a lot or a l i t t l e ) . Sinclair suggests that the precise and regular use of relative expressions reflects the same a b i l i t i e s as the operations of seriation and conservation of quantity, i.e., decentra-tion and the capacity to coordinate. Harasym, Boersma and Maguire (1971) maintain that children at the preconservation level do not differentiate the meanings of more and less on the basis of their similar profiles for the two words on the semantic differential. An obstacle in applying their data to the 'less = more' 16 phenomenon is that the non-conservers in their study made associations for both more and less which the conservers made only for less. If this, then, is interpreted to indicate the attribution of one meaning to the two words, i t would be the meaning of less, rather than that of more as suggested by Palermo (1973) and Clark (1973). Although the spontaneous production of relative terms was clearly correlated with the development of conservation in Sinclair's study, tasks which required the comprehension of these terms were carried out equally well by children at preoperational and concrete operational levels. However, the studies cited above have demonstrated that young children do have d i f f i c u l t y with some tasks which require the comprehen-sion of comparative terms. If the correlation observed by Sinclair be-tween conservation and the spontaneous production of comparatives in the description of inequalities is interpreted as an indication of a relation-ship between cognitive capacity and verbal expression, then at an earlier stage a cognitive correlate for the development of the comprehension of the same terms might also be demonstrable. Klatsky, Clark and Macken (1973) present evidence that the apparent asymmetry in the acquisition of polar adjectives is based on differences in cognitive complexity, rather than on the more frequent occurrence of positive polars in the language. Their subjects were taught to use nonsense syllables for positive and negative, adjectives describing size, height, length and thickness. For each negative term, more t r i a l s were required to reach criterion, and more errors were made on post-criterion tasks than for the corresponding positive term. They 17 then hypothesized that comparative judgments of dimension are made in terms of two primary reference points: (1) the origin, which is common to both stimuli, and (2) the secondary reference point, or endpoint of the standard, which becomes the origin for the item to be compared. Thus, consideration of an item with less extension than the standard requires a 'doubling back' or reversal in direction from the endpoint of the standard to the endpoint of the item of comparison. The consideration of endpoints in the evaluation of quantity has also been discussed by Piaget (1968). Before children have an opera-tional understanding of quantity which is independent of changes in shape and rearrangement of components, judgments are based on the endpoints of visual regard. According to Piaget, Inhelder and Szeminska (I960), the earliest judgments of length are dependent on primitive topological 'intuitions' of order. The understanding of length proceeds from a stage at which children do not yet concern themselves with intervals between endpoints. Consequently, the child makes judgments concerning the lengths of objects in terms of the order or positions of their boun-daries. The reliance by children on order of furthermost extremities in comparing lengths is demonstrated when preoperational children judge a stick as longer or shorter than another of identical length when i t is displaced horizontally. Preoperational children, then, would not be expected to make comparisons of extension in terms of two endpoints, as the analysis of Klatsky ejt a_l. (1973) suggests. 1.13 Bases for the Apparent Confusion of more and less E. Clark (1973) has used data gathered from research on the •V 18 temporal markers, before and after, dimensional adjectives such as longer and shorter, and the quantitative terms more and less as evidence for the earlier acquisition of positive than of negative polars postulated by the 'Semantic Feature Hypothesis'. The data which she used in support of the earlier acquisition of before than after were attained in a comprehension task (Clark, 1971). However, the design of her experiment failed to separate the variable of a b i l i t y to deal with complex sentence structures from the variable of comprehension of before and after. Sentences such as 'Before (after) he jumped the gate, he patted the dog' were given to children between the ages of 3;0 and 5;0 to be acted out. The results showed higher comprehension scores for sentences using the word before than for those using the word after. The data could also be interpreted in terms of the sentence structure of the instructions: In a l l 'after' sentences, the f i r s t event was either described second or was contained in the subordinate clause, while half of the 'before' sentences gave the f i r s t event in the f i r s t clause which was also the main clause of the sentence (cf. Amidon and Carey, 1972). The research by Donaldson and Wales (1970), which reports a difference in comprehension scores in favor of the positive-pole terms is basic to Clark's reference to the acquisition of comparative dimen-sional terms as support for the 'Semantic Feature Hypothesis'. In view of the finding of Eilers et a l . (1974) that children in the 2;6 to 3;6 age range exhibit a significant bias in favor of negative dimensional terms, the propriety of including dimensional terms as evidence for such a hypothesis is also questionable. Since the body of data relating to the acquisition of more and less is also inconclusive (see Table 1 at the end of this section), there is no firm foundation for the generalization proposed by Clark that positive or unmarked adjectives are understood before their negative, or marked counterparts. There may exist a basis on which terms which refer to variables of time, space and number may be associated. Piaget (1952) has shown that number is understood only in terms of 'intuitive' impressions. Comparisons of quantity are made on the basis of the spatial organiza-tion of the members of the set. Two lines of objects which are equal in number, for example, are judged as unequal when the elements of one row are either spread out or crowded together. In this case, terms which are. used in comparisons of quantity would reflect a similar level of development as those used in comparing length; e.g., more would have essentially the same criterion for application as would longer or denser. Ferreiro and Sinclair (1971) related the development of the a b i l i t y to express a reciprocal temporal relationship between two events to the development of the re v e r s i b i l i t y which characterizes operational thinking. Preoperational children described a sequence of two events separately and in random fashion, using neither relational clauses nor adverbial expressions, such as then, or afterwards. Children at an intermediate level always described the two events in the order in which they occurred, and often linked them with adverbial expressions. These children were unable, when asked, to describe the series of events be-ginning with the (temporally) second one. Children who had reached the level of concrete operations were able to describe the two events using subordinate clauses and to describe the events in reverse order while maintaining the correct temporal relationship. Ferreiro and Sinclair propose that children in the f i r s t cate-gory specified above regard the two events as two separate and unrelated entities, neither of which is a reference point for the other. This is interpreted as an indication of egocentrism and lack of cognitive decen-tration. At the intermediate level, the temporal link between the two events is strongly established, but the dependency has only one direction. It is not u n t i l the development of reversibility that the relationship becomes reciprocal. They suggest that this interpretation is not one which is limited to temporal relationships, but one which may be used as a guideline for a more general explanation of the development of some aspects of language. The comprehension of labels for quantitative relationships and dimensional comparatives seems to follow a course of development which is similar to the use of labels for temporal relationships. In identi-fying the member of a pair which has more, and the one which has less, very young children appear to make random choices. At a later, stage, children seem to order objects in one direction only and are unable to deduce from (A greater than B) that (B less than A). The a b i l i t y to shift direction in considering an inequality (i.e., to consider A's relation to B, as well as B's relation to A) re-quires mobility of thought, which Piaget and Inhelder (1956) have dis-cussed in relation to the development of the a b i l i t y to perceive and to reproduce linear order. The abstraction of order takes place as a result of increasing coordination of actions such as transferring or transporting elements mentally. The i n i t i a l construction of order requires the co-ordination of eye movements, a sensorimotor activity without which order cannot even be perceived. According to Piaget and Inhelder (1956), the simple relations of [A next to B] and [B next to C] must be coordinated in order for the series of ABC to be reproduced. The child must be able to distinguish between the two directions of travel in considering a series, and to maintain a constant direction while transferring elements from the model to the copy. Similarly, in order to understand the reciprocal relation-ship between unequal quantities the child must be able to coordinate the two relations implicit in considering a pair of inequalities, and to dis-tinguish between the two directions of comparison. The intermediate stage described by Ferreiro and Sinclair in the comprehension of temporal relationships, in which the temporal link has only one direction may be paralleled in the development of the comprehen-sion of quantitative relationships by a stage in which the quantitative comparison is unidirectional. Such a stage would account for the observed ina b i l i t y of some children to identify one of the poles of an asymmetri-cal pair. " Ferreiro and Sinclair relate the development of the comprehension of reciprocal temporal relationships to the development of reversibility of thought, as indicated by the a b i l i t y to conserve. The comprehension of reciprocal quantitative relationships, then, might also be dependent on the development of reve r s i b i l i t y . Comparisons between quantities in-volve the consideration of spatial relations, since the perceptual cues are length and density. Thus, the development of the a b i l i t y to coordi-nate the more primitive spatial relation of proximity as indicated by the a b i l i t y to reproduce a linear series would be expected to precede the a b i l i t y to coordinate comparative relations of length. Piaget and Inhelder (1956) outline three stages in the develop-ment of linear ordering: Stage I: inab i l i t y to reproduce a sequence, even when the copy is directly beneath the model; Stage Ila : a b i l i t y to make a copy directly beneath the model; Stage l i b : a b i l i t y to copy when there is a slight offset between the model and the copy; and Stage III: a b i l i t y to make a copy in reverse order to the model. The transition from Stage Ila to Stage l i b is purported to be the result of the development of the beginnings of mobility in motor coordination. The coordination of relations of proximity which is required for the operation of linear ordering is one which would precede the coordina tion of geometrical relations, such as length i f , as Piaget maintains, topological organization precedes geometrical organization of space. It would thus*be expected that sensorimotor mobility, as demonstrated by the a b i l i t y to reproduce linear order, would precede the understanding of reciprocal geometrical relationships. If development in the a b i l i t y to handle instructions containing the words more and less is related to cognitive development, the varia-tions among studies which report different results in assessing the comprehension of more and less would be attributable to differences in the cognitive demands of the tasks employed. A preliminary suggestion is that those tasks which do not specify a standard for comparison require greater mobility of thought than those which do. In the latter type of task, the child is asked to determine the direction in which the object of comparison differs from the standard. This does not require coordina-tion between the relations more than and less than, whereas the simul-taneous comprehension of both relations which exist between a pair of inequalities requires the development of reve r s i b i l i t y to the extent that the child is able to use the same item alternately as a standard and as an item of comparison. 1.14 The Experimental Assessment of Comprehension of more and less; A Comparison of Methodologies The various conditions under which the comprehension of more and less have been assessed are outlined in Table 1 at the end of this section. Palermo's (1974) survey of types of stimuli used for experi-mental comparison indicated that responses to the word less are not significantly different for judgments of continuous vs discontinuous quantities. A l l of the tasks using discontinuous quantities presented by Palermo compared two sets of objects in one-to-one correspondence. The comparisons of quantities not in isomorphic correspondence carried out by Mehler and Bever (1968) did not investigate the comprehension of the word less, but indicated only that the criterion for the assignment of the word more to a quantity of objects appears to develop toward one which is based on linear comparisons between the ages of two and four years. Donaldson and McGarrigle (1974) also showed that when a criterion with a topological basis (such as fullness) is available, some children s h i f t from linear to topological evaluation. Past investigations have varied in the response type required of the child. The situations differed as to whether subjects were required (a) to produce more and less, (b) to identify the appropriate quantity in response to the experimenter's production of the terms, or (c) to create inequalities as directed by the experimenter; and within these categories there are further variations. In producing the terms, c h i l -dren were required to describe a pair of unequal quantities (Sinclair, 1969; Griffiths e_t al_. , 1967) or to answer a question such as 'Does this one have more or less than that one?' (Griffiths e_t al_., 1967). In creating inequalities, children were asked either to put more and less objects in specified places: e.g., 'Give more to the boy and less to the g i r l ' (Inhelder, Sinclair and Bovet, 1974), or simply to make i t so that either of two sets had more (or less) than the other (e.g., 'Put more (less) apples on one tree than the other') and then to point out the one with more and the one with less (Donaldson and Balfour, 1968; Palermo, 1973). Results for the word more are consistent for a l l tasks which present continuous quantities or quantities in one-to-one correspondence. Except for the 2-year-old subjects in Weiner's (1974) study, consistently correct responses were given in dealing with more regardless of the task. Responses to the word less, however, do seem to be affected by the nature of the task. In making judgments of static inequalities, children gave responses which indicated comprehension of the word less when they were required to choose between the words more, less, and the same in relating a quantity to a specific standard (Griffiths, Shantz and Sigel, 1967). However, when subjects were asked to choose which of a pair of quantities had less, performance was at chance level or below (Donaldson and Balfour, 1968; Palermo, 1973; Weiner, 1974). In creating inequali-ties, children were able to follow instructions which directed them to put less in one place than another, but following this many were unable to correctly identify the one which had less (Donaldson and Balfour, 1968; Palermo, 1973). When i t was specified which member of a pair was to re-ceive more and which was to receive less, children's responses were consistently correct. In this situation, comprehension of the word more alone might be sufficient for correct responses to the instructions. However, the conclusions drawn by Donaldson and Balfour (1968) and Palermo (1973, 1974) that less is understood as more is not supported by Inhelder, Sinclair and Bovet's (1974) findings. If this were the case, the instructions would be understood as 'Give more to the g i r l and more to the boy', the result of which would have been equal distributions or randomly varying inequalities. The age groups overlap sufficiently between Inhelder, Sinclair and Bovet's (1974) and Palermo's (1973) subjects to warrant the conclusion that age differences were not totally responsible for the differences in results. Even among Palermo's 7-year-old subjects, six of the 32 children performed as i f less were understood to be synonymous with more. Thus, i t would be expected that at least some of Inhelder, Sinclair and Bovet's 4.5- to 8-year-old subjects would have had d i f f i c u l t y with the word less had they been given a task similar to the one presented by Donaldson and Balfour (1968) and Palermo (1973). The tasks given by Donaldson and Balfour (1968), Palermo (1973, 1974) and Weiner (1974) required that both stimuli be compared with each other. The child was shown two stimuli and asked 'Which one has more (less)?'. In this case, neither member of the pair is named as a stan-dard, and the comparison must therefore be made in two directions: A has more because B has less, and vice versa. Each stimulus must serve as the standard of comparison for the other i f the reciprocal relation-ship is to be understood. Gr i f f i t h s , Shantz and Sigel's (1967) study, showing a high rate of correct responses to more and less questions, used a task in which children were given one stimulus as a standard and were then presented with a second stimulus, which they were required to.judge as more, less, or the same. The child was not required to define a relationship between two objects, but only to make a unidirectional comparison from the stan-dard to the item in question. Another situation in which children responded consistently correctly to the words more and less is the one created by Inhelder e_t a l . (1974), in which both terms are used in the instructions, and i t is not necessary to infer one relation from another. The child is given both relations; e.g., more marbles for the g i r l and less for the boy. There is evidence to suggest that under some conditions children demonstrate the a b i l i t y to differentiate more and less on both produc-tion and comprehension tasks. The situation in which the greatest confu-sion of these terms is apparent is the one in which the child is required to choose the one with more or less from a pair of unequal stimuli. 27 Perhaps, then, i t is some aspect of this type of task, rather than the words more and less which causes d i f f i c u l t y for some children. Insert Table 1 about here. Table 1. Summary of Previous Research by Task C l a s s i f i c a t i o n Type of Task Inves tigator(s) S_'s Ages N Results Producing S p e c i f i e d Inhelder, 4;6-8;0 88 more: A l l Ss except one Inequalities S i n c l a i r less: (e.g., give more to and Bovet longer: c a r r i e d out a l l i n -A and less to B) (1974) shorter: thicker: structions c o r r e c t l y thinner: Producing Unspecified Donaldson 3;5-4;l 15 more: 14 S_s made the 2 Inequalities and quantities unequal (e.g., put more or Balfour less on one than (1968) le s s : 13 S_s made the 2 another) quantities unequal I d e n t i f y i n g Inequalities Weiner 2;l-3;6 16 more: 47% correct i n one-to-one (1974) le s s : 41% correct correspondence ( S t a t i c , l i n e a r ) 2;l-3;2 ' 16 more: 56%, correct le s s : 44%, correct 3;5-3;ll 18 more; 94%. correct le s s : 49% correct 3;0-4;4 36 more: 83% correct le s s : 69% correct to * CO I Table 1 (Cont'd) Type of Task Investigator(s) S_'s Ages N Results Identifying Inequalities in one-to-one corres-pondence Donaldson and Balfour ( 1 9 6 8 ) 3 ; 5 - 4 ; l 1 5 more: less: 9 1 % 277. correct correct (Static, non-linear) Palermo ( 1 9 7 3 ) 3 and 4 X = 3 . 9 1 6 more: less: 947o 287o correct correct 4 and 5 X = 4 . 8 1 6 more: less: 987o 457o correct correct 5 - 6 (Kindergarten) 3 2 more: less: 957o 1 6 % correct correct 6 - 7 (Grade 1 ) 3 2 more: less: 9 7 % 5 2 % correct correct 7 - 8 (Grade 2 ) 3 2 more: less: 1 0 0 % correct 9 1 % correct Identifying Inequalities in one-to-one correspondence (Transformational, 1inear) Weiner ( 1 9 7 4 ) 2 ; l - 3 ; 6 2 ; l - 3 ; 2 1 6 1 6 more: less: more: less: 5 2 % 5 5 % 6 2 % 5 0 % correct correct correct correct 3 ; 5 - 3 ; l l 1 8 more: less: 9 9 % 5 3 % correct correct 3 ; 0 - 4 ; 4 3 6 more: less: 8 6 % 7 1 % correct correct Table 1 (Cont'd) Type of Task Investigator(s) S_'s Ages N Results Identifying Inequalities Donaldson and 3 ; 5 - 4 ; 1 1 5 more: 8 4 % correct in one-to-one corres- Balfour ( 1 9 6 8 ) less: 2 9 % correct pondence (Transformational, non-linear) Palermo 3 and 4 1 6 more: 9 7 % correct ( 1 9 7 3 ) 1=3.9 less: 2 2 % correct 4 and 5 1 6 more: 977o correct X = 4 . 8 less: 507o correct 5 - 6 3 2 more: 9 4 % correct (Kindergarten) less: 9% correct 6 - 7 3 2 more: 1 0 0 % correct (Grade 1 ) less: 5 6 % correct 7 - 8 3 2 more: 1 0 0 % correct (Grade 2 ) less: 8 8 % correct Identifying Inequalities Bever, 2 ; 0 - 2 ; 5 1 4 more; 9 3 % correct not in one-to-one Mehler and correspondence Epstein ( 1 9 6 8 ) 2 ; 6 - 2 ; 1 1 1 9 more: 3 7 % correct (Static) 3 ; 0 - 3 ; 5 2 8 more: 3 8 % correct 3 ; 6 - 3 ; 1 1 3 0 more: 5 7 % correct 4 ; 0 - 4 ; 1 1 5 0 more: 5 0 % correct Table 1 (Cont'd) Type of Task Investigator(s) S_'s Ages N Results Identifying Inequalities not in one-to-one Piaget ( 1 9 6 8 ) 2 ; 3 - 3 ; 1 0 2 9 more: 507o correct correspondence (Trans formational) Mehler and Bever ( 1 9 6 7 ) 2 ; 4 - 2 ; 7 2 ; 8 - 2 ; l l 2 2 2 9 more: more: 1 0 0 7 , correct 8 0 7 , correct 3 ; 0 - 3 ; 3 3 0 more: 5 7 7 , correct 3 ; 4 - 3 ; 7 3 6 more: 5 6 7 , correct 3 ; 8 - 3 ; l l 1 3 more: 1 5 7 , correct 4 ; 0 - 4 ; 3 2 2 more: 2 0 7 , correct 4 ; 4 - 4 ; 7 2 6 more: 7 0 7 , correct Spontaenous Production Sinclair ( 1 9 6 9 ) Not given Not given Ss with conservation: 1 0 0 7 , used comparative terms Ss without conservation: 907o used absolute terms El i c i t e d Production Griffiths, Shantz and Sigel ( 1 9 6 7 ) 4 ; 1 - 5 ; 2 5 4 more: less: 7 2 7 , correct 8 0 7 , correct 32 a 1.2 Summary and Statement of the Problem The performance of children on tasks which deal with quantitative or dimensional polar terms under varying conditions has not been consis-tent either in showing that these terms are indeed problematic or in demonstrating reliable response patterns. Consequently, attempts to explain from a semantic viewpoint responses to instructions which contain such terms are premature. E. Clark's (1973) explanation of the response bias in favor of positive polar adjectives, as demonstrated by Donaldson and Balfour's (1968) and Palermo's (1973) research, is not supported given the absence, of this bias in Griffit h s , Shantz and Sigel's (1967) study, and given the bias in favor of the opposite pole found by Eilers, Oiler and Ellington (1974). The quantitative terms, more and less, dimensional adjectives, such as longer and shorter, and the temporal markers, before and after, were grouped together by E. Clark (1973) for the purpose of demonstrat-ing common features in their acquisition which would support the 'Semantic Feature Hypothesis'. She postulates that the meanings of positive polar adjectives are acquired before their negative counter-parts, and that at some stage in development, the negative and positive members of a pair are both understood to have the meaning of the positive pole. The contradictory results of the various studies which have dealt with children's understanding of quantitative and dimensional compara-tive adjectives and the inadequacy of the research regarding before and after, suggest that such a hypothesis is tenuous at best. One may, on the other hand, find j u s t i f i c a t i o n in the literature for seeking common bases for children's d i f f i c u l t i e s with spatial, temporal and quantitative terms. Both linguistic (H. Clark, 1973) and cognitive (Piaget, 1969) evidence suggests that the organization of time is based on that of space. Perhaps Ferreiro and Sinclair's (1971) finding, that the a b i l i t y to represent temporal order is preceded by the development of cognitive decentration and the reduction of egocentrism, is indicative of a more general principle, which may be applicable to dimensional terms as well. Moreover, i f considerations of quantity are based on the geometrical variables of length or area (cf. Piaget, 1952), one would expect the acquisition of terms referring to quantity, such as more and less, to be related to the acquisition of dimensional terms. The variation among studies in the requirements of the tasks presented precludes the abstraction of definitive generalizations from existing research on relational terms (see Table 1). A possible cogni-tive basis for children's performance in situations in which they appear to confuse these terms is examined in the present investigation, and performances on two types of tasks which have given conflicting results in previous research are compared in order to determine whether the particular conditions of the task are a significant factor in the presence and degree of confusion of comparative terms. Piaget and Inhelder (1956) trace the development of the percep-tion of relationships by preoperational children on a task which deals with the topological relationship of proximity. The coordination of relations of proximity leads to the perception and reconstruction of linear order. The inability of the preoperational child, whose thinking 34 -•* is s t i l l in large measure sensorimotor, to distinguish between two direc-tions of travel in reconstructing a linear series or to maintain a con-stant direction is attributed to lack of reve r s i b i l i t y in perceptual motor activity. The concept of linear order is acquired as a result of recon-struction through ordered actions. The development of the perceptual motor re v e r s i b i l i t y which leads to the perception of a series of rela-tions of proximity, as indicated by a b i l i t y to reconstruct linear order, is a. cognitive capacity which might also underlie the a b i l i t y to reverse direction of comparison when dealing with asymmetrical relations among objects. If the development of reve r s i b i l i t y of thought is preceded by a stage of semi-logic (Piaget et_ a_l., 1968) when, in relationships of order, the dependency has only one direction (Ferreiro and Sinclair, 1971), the comprehension of the reciprocal relationship between a pair of in-equalities may be preceded by a comprehension of relations of quantity which is unidirectional. Such a stage could account for the response biases which have been observed by some authors (Donaldson and Balfour, 1968; Palermo, 1973; Eilers, Oiler and Ellington, 1974). This unidirec-tionality would be reflected in the inability to perceive and reconstruct relationships of proximity, as indicated by the inability to reconstruct linear order, i f a common non-reversibility of thought is the basis for both. The aims of the present investigation are as follows: (1) To determine whether the contradictory and inconsistent results observed in previous investigations are attributable to dif f e r -35 ences in experimental design; this w i l l be carried out through examination and comparison of children's responses on (a) the task of choosing appropriate test items in response to the experimenter's production of the words more and less, longer and shorter, and (b) the task of choosing between and using such verbal labels appropriately when presented with test items in relation to given standards. (2) To determine whether a relationship is demonstrable between perceptual motor i r r e v e r s i b i l i t y (as indicated by an inab i l i t y of the child to reconstruct linear order) and d i f f i c u l t y in making comparisons (as evidenced by a response bias when choos-ing between pairs of unequal quantities). Performance on Piaget and Inhelder's (1956) linear ordering task w i l l be com-pared with performance on the stimulus-choice task, as in 1(a) above. (3) To examine the notion of 'fullness' in the earliest comprehen-sion of more, through the comparison of children's tendencies to assign that verbal label to the correct member of a stimulus pair, (a) when that set contains empty spaces and (b) when a l l available spaces in that set are occupied. CHAPTER 2 EXPERIMENT I 2.1 Method 2.11 Subjects Subjects selected for inclusion in this study were eight g i r l s and six boys between the ages of 2;6 and 4; 6, with a mean age of 3 years, 5 months. 2.12 Materials Stimuli used to assess comprehension of more and less were three pairs of 12 x 2 inch cardboard strips on which were glued unequal numbers of candies (M&M's). The f i r s t three pairs illustrated in Figure 1 were used in Experiment I. The paired comparisons were: 7 and 3, 7 and 5, and 5 and 3. For each comparison, the two sets of candies were the same color, and candies were regularly spaced and always in visual one-to-one correspondence (i.e., each candy in the row with 'less' occupied a space which was directly beneath a space in the row with 'more'). The maximum number of candies which could be placed on the strips was 7. For two of the comparisons, a l l of the spaces were f i l l e d on the strip which had 'more' candies. Insert Figure 1 about here. 36 37 •o o o o o o o Pair 1 o o o o o o o . o o o Pair 2 o o o o o o o o o o Pair 3 o o o o o o o 4 o o Figure 1. Stimuli for more/less Tasks in Experiment I (Pairs 1-3) and for more/less Type A Tasks in Experiment II (Pairs l-ty. 38 The comprehension of longer and shorter was tested using card-board strips 0.5 inch in width and 2, 4 and 6 inches in length. The task of reconstructing linear order was carried out (a) with a chain of five plastic beads which varied in shape and color, and (b) with a chain of seven separated beads, five of which were identical in shape and color to beads in the model chain. 2.13 Procedure Each of the following tasks were administered in random order to each subject: Task M/L-l: Identification of more and less. Subjects were shown two rows of M&M's. Each of three pairs was presented twice, once with the question 'Which one has more candies?', and once with the question 'Which one has less candies?'. The order of presenta-tion and the order of the two questions were randomized, with the excep-tion that one more question and one less question were asked for each set, and the two presentations of the same set were never consecutive. Task M/L-2; Production of more and less. The same three pairs were presented. Again, each pair was presented twice, but not consecutively, and with this exception, the order of presentation was random. For each pair, the subject was given one row and told to 'look at these candies'. The other row was then presented with the question 'Does this one have less candies or more candies?', or 'Does this one have more candies or less candies?'. For each of the three paired comparisons, each set was presented once as the standard and once as the item of comparison, thus giving a total of six comparisons. Task L/S-l: Identification of longer and shorter. The subject was shown two cardboard strips of equal width and different lengths, and was asked 'Which one is longer?', or 'Which one is shorter? Order of presentation and questions was again randomized, with the same exceptions noted for the more/less tasks. The lengths compared were: 2 inches and 4 inches, 4 inches and 6 inches, and 2 inches and 6 inches. For each of the more/less tasks the two stimuli were placed directly in front of the subject, one above the other with the edges of the strips aligned, as shown in Figure 1. For the longer/shorter task, the l e f t edges of the cardboard strips were aligned. The relative place ment of the two stimuli varied randomly. Linear Ordering: The procedure described by Piaget and Inhelder (1956) was followed. The subject was shown the model chain of five beads and was asked to find the matching bead for each of the beads on the chain, one at a time. When i t was clear that the subject was able to match beads, the experimenter pointed out the model chain and told the subject to make another one just like i t , emphasizing the order of the beads. Subjects who were successful on this task were then asked to repeat what they had done to construct the second chain beside the model. If this was carried out correctly, the subject was asked to copy the model again but to reverse the order. There were no standard instructions for this task; the experimenter explained and demonstrated the task un t i l she fel that the subject understood as well as possible what was expected. 40 2.14 Scoring Tasks M/L-l and M/L-2; Each correct identification or production of more and less was given one point. The maximum score per subject for each more/ less task was 3 points for more and 3 points for less. Task L/S-l; Each correct identification of longer and shorter was given one point. The maximum score per subject was 3 points for longer and 3 points for shorter. Linear Ordering: Behavior on the linear ordering task was cla s s i f i e d according to Piaget and Inhelder's (1956) c r i t e r i a as follows: Stage 0: The subject f a i l s to choose five beads which are the same as those in the model chain; Stage I: The subject chooses the same five beads, but does not place them in the same order as the model chain; Stage Ila: The order is reproduced correctly when the copy is directly beneath the model but not when the copy is beside the model; Stage l i b ; The subject reproduces the order correctly when the copy is beside the model, but cannot reverse the order; and Stage III:. The subject successfully makes a copy in reverse order to the model. 2.2 Results 2.21 More vs less; longer vs shorter Table 2 shows mean scores and percent correct for the positive and negative terms on the three comparison tasks. A Wilcoxon Matched-Pairs Signed-Ranks Test was applied to each pair of scores. Although the mean scores are consistently higher for the negative member of each pair, there are no significant differences between the scores for more and less in either task, nor between the scores for longer and shorter. None of the 1 4 children in Experiment I responded consistently to the word less as i f i t meant more; i.e., there were no cases of 1007o correct identifications of more and 1 0 0 7 . incorrect identifications of less. Only three of the subjects had a difference of more than one point between their scores for more and less on Task M/L-l. In each case, the higher score was for less. Insert Tables 2 and 3 about here. 2 . 2 2 Age Effects Percent correct for four age categories on more/less-1, more/  l e s s - 2 , and longer/shorter -1 tasks are presented in Table 3 . The scores of the youngest age group are generally lower than those of the others. With the exception of l e s s - 2 and shorter - 1 , the scores for this group are near chance level. These data suggest that the subjects below the age of 3 ; 0 did not understand the comparative words involved in any of the tasks. Subjects above the age of 3 ; 0 obtained scores above chance level for each word on a l l tasks, indicating some comprehension of each of the words tested. There are no apparent age trends in the scores for the three older categories. However, due to small sample sizes, particularly in the 3 ; 6 - 3 ; 1 1 age range, the possibility of age differences within the 3 ; 0 42 TABLE 2 Experiment I. Mean scores and percent correct for more/less and longer/ shorter tasks (maximum score = 3 for each term). N = 14. Task Positive Term Negative Term Mean Percent Mean Percent M/L-l 1.93 64 2.43 81 M/L-2 2.07 69 2.50 83 L/S-l 1.86 62 2.43 81 TABLE 3 Experiment I. Correct responses (7„) on Tasks M/L-l, M/L-2 and L/S-l for four age groups. Age Group N 2;6-2;ll 5 3;0-3;5 4 3;6-3;ll 1 4;0-4;6 4 more-1 46.7 75.0 100.0 66.7 less-1 53.3 91.7 100.0 100.0 more-2 60.0 66.7 100.0 75.0 less-2 86.7 75.0 100.0 91.7 longer-1 46.7 66.7 100.0 66.7 shorter-1 66.7 100.0 100.0 75.0 to 4; 6 range may not be rejected. 2.23 Stimulus Effects Table 4 shows the number and percentage of subjects who chose the correct stimulus in response to M/L-l questions and who gave the correct response to M/L-2 questions for each stimulus. The number of children who labelled the larger set as more (Task M/L-2) for Pair 3 is consider-ably smaller than the number who did so for Pairs 1 and 2. A similar tendency is evident in the scores for Task M/L-l. For Pair 3, the least number of correct responses was obtained to requests for the one with more. The sets which contained more candies in Pairs 1 and 2 were f u l l ; i.e., the candies extended across the entire length of the cardboard strip. According to the same criterion the more numerous set of Pair 3 was not f u l l , since the row of candies covered only part of the strip (see Figure 1). It is apparent, then, that 'fullness' is used by some children as a criterion for application of the word more. Insert Table 4 about here. 2.24 Summary None of the children responded to the word less as i f i t were synonymous with more. In fact, the difference between scores for the positive and negative members of each word pair consistently favored 44 TABLE 4 Experiment I. Number of and percentage correct responses for each stimulus on Tasks M/L-l and M/L-2. N = 14. Pai r Stimulus Task M/L' -1 Task M/L--2 No. Correct Percent No. Correct Percent 1 more (7) 7 65.7 11 78.6 less (5) 11 78.6 11 78.6 2 more (7) 10 71.4 12 85.7 less (3) 13 92.9 12 85.7 3 more (5) 8 57.1 5 35.7 less (3) 10 71.4 11 78.6 45 • li-the negative term. The data indicate that children under the age of 3;0 do not make consistent responses to the words more, and less and longer and shorter, and that for some children the attribute of 'fullness' plays a role in their understanding of the word more. CHAPTER 3 EXPERIMENT II 3.0 Introduction Since the expected positive response bias was not observed in any of the subjects of Experiment I, a further 16 children were tested using the same stimuli plus a fourth pair for the more/'less tasks (see Figure 1) and a fourth comparison for the length tasks. Four additional pairs of stimuli were presented for more/less comparisons. These differed from the others in that for each pair, the lengths of the two rows were the same (see Figure 2). These stimuli were added in order to determine the impor-tance of length as a factor in the subjects' a b i l i t y to make more/less comparisons. 3.1 Method 3.11 Subjects Six boys and ten g i r l s between the ages of 2;10 and 4;2, with a mean age of 3;6 were selected as subjects. 3.12 Materials Stimuli were those used in Experiment I with the addition of one paired comparison of 4 and 2 candies (see Figure 1) for the more/less tasks and the addition of one cardboard strip 8 inches in length for the longer/shorter tasks. Four additional pairs of inequalities were con-structed. Again, M&M's were attached to cardboard strips, but in this 46 47 case the rows were the same length and the row with less candies was more widely spaced. The numbers compared were the same as those used i n the equally spaced rows. These s t i m u l i are i l l u s t r a t e d i n Figure 2. Insert Figure 2 about here. 3.13 Procedure The questioning procedure followed that of Experiment I. There were eight comparisons for more and l e s s : four with the s t i m u l i i l l u s -t r a ted i n Figure 1 (Type A) and four with the s t i m u l i i l l u s t r a t e d i n Figure 2 (Type B) , and four comparisons for longer and shorter. Each subject received a l l of the following tasks: Task M/L-l(A): I d e n t i f i c a t i o n of more and less with Type A s t i m u l i , as i n Experiment I, Task M/L-l. Task M/L-1(B): I d e n t i f i c a t i o n of more and less with Type B s t i m u l i . Task M/L-2(A): Production of more and less with Type A s t i m u l i , as i n Experiment I, Task M/L-2. Task M/L-2(B): Production of more and less with Type B s t i m u l i . Task L / S - l ; I d e n t i f i c a t i o n of longer and shorter, as i n Experiment I, Task L / S - l . Task L/S-2; Production of longer and shorter. 48 o o O 0 o o o. Pair 1 o o o o o o o o o o Pair 2 o o •o o o o o o o 0 Pair 3 o o o o o o o -Pair 4 o o Figure 2. Stimuli for more/les s_ Type B Tasks in Experiment II. 49 v Linear Ordering: As in Experiment I. 3.14 Scoring The scoring procedure was the same as that of Experiment I. The maximum scores were 4 points each for the positive and negative terms in each comparison task. 3.2 Results 3.21 More vs less; longer vs shorter Table 5 presents mean scores on Tasks M/L-l and M/L-2 for both Type A (equal length) and Type B (unequal length) stimuli and on Tasks L/S-l and L/S-2. Since subjects' response patterns were similar for the two types of stimuli on both more/less tasks, the scores for Type A and Type B stimuli were combined for analysis. The number of correct responses to the question 'Which has more?1 is significantly greater than the number of correct responses to the corresponding 'less' question (Task M/L-l: Wilcoxpn T = 7, N s_ r = 10, p < .05). The difference between the number of correct productions of more and less (i.e., Task M/L-2) is not significant, nor is the difference between the number of correct productions of longer and shorter. The scores for identification of longer and shorter (Task L/S-l) are identical. Only three subjects demonstrated a positive response bias on Task M/L-l. With a maximum score of 8 for each term, these three children each had 2 or fewer errors for more-1 and 6 or more errors for less-1. A fourth subject showed such a response pattern only for the Type B stimuli: 50 with a maximum scora of 4 per term, she made 4 errors for less and none for more. This subject showed a negative response bias on Task L/S-l: with a maximum score of 4 per term, she made no errors for shorter and 4 for longer. There were no children who showed such an asymmetry of scores for the positive and negative terms on Task M/L-2 or L/S-2. One subject refused to answer any of the questions on these two tasks. The difference in mean scores between Tasks M/L-l and M/L-2 indicate that these two tasks are not equivalent assessments of children's comprehension of more and less. Insert Tables 5 to 7 about here. 3.22 Age Effects Table 6 shows percent correct on each task for four age cate-gories. As in Experiment I, a l l scores for subjects over.the age of 3;0 were above chance level, indicating comprehension of each of the words tested. Only one subject was below the age of 3;0, and this subject was one of the three who showed a positive response bias on Task M/L-l. 3.23 Stimulus Effects The number and percentage of correct responses to each of the four Type A and four Type B stimuli are given in Table 7. There are no large differences in the numbers of correct responses to any of the stimuli. 51 TABLE 5 Experiment II. Mean scores and percent correct for Type A and B stimuli for more/less tasks (maximum score for each stimulus type = 4 for each term) and for longer/shorter tasks (maximum score = 4 for each term). N = 16. Task Stimulus Type Positive Term Mean Percent Negative Term Mean Percent M/L-l M/L-2 L/S-l L/S-2 A B Total A B Total Total Total 3.44 3.38 6.82 3.13 3.00 6.13 3.31 3.13 86.0 84.2 85.1 78.2 75.0 76.6 82.8 78.2 2.44 2.50 4.94 3.33 3.47 6.80 3.31 3.53 61.0 62.5 61.8 83.2 86.8 85.0 82.8 88.2 52 TABLE 6 Experiment II. Correct responses (%) on Tasks M/L-l, M/L-2 and L/S-l for four age groups. Age Group N 2;G-2;11 1 3;0-3;5 4 3;6-3;ll 10 4;0-4;6 1 more-1 100.0 68.8 92.5 100.0 less-1 25.0 71.9 66.2 87.5 more-2 25.0 75.0 81.2 87.5 less-2 50.0 70.8 91.2 100.0 longer-1 75.0 81.2 90.0 100.0 shorter-1 50.0 75.0 95.0 100.0 longer-2 50.0 75.0 82.5 100.0 shorter-2 50.0 83.3 92.5 100.0 53 TABLE 7 Experiment II. Number of and percentage correct responses for each stimulus on Tasks M/L-l (N = 16) and Task M/L-2 (N =15). Task M/L-l Task M/L-2 Pair Stimulus No. Correct Percent No. Correct Percent more(7) A 14 87.5 12 80.0 B 15 93.8 13 86.7 less(5) A 11 68.8 12 80.0 B 10 62.5 13 86.7 more(7) A 13 81.2 11 73.3 B 13 81.2 11 73.3 less(3) A 11 68.8 14 93.3 B 14 87.5 10 66.7 more(5) A 14 87.5 10 66.7 B 14 87.5 10 66.7 less(3) A 9 56.2 13 86.7 B 8 50.0 12 80.0 more(4) A 16 100.0 12 80.0 B 13 81.2 11 73.3 less(2) A 12 75.0 13 86.7 B 10 62.5 15 100.0 54 /» 3.3 Combined Results of Experiments I and II 3.30 Introduction For the remaining analyses the scores of the children in Experi-ment I and Experiment II were combined. Only the stimuli which were common to both groups are considered. Since the only difference in pro-cedure between the two experiments was the addition of some stimuli, and since both groups of children were drawn from the same population, i t was considered justifiable to combine the results of the experiments in order to obtain clearer indications of the effects of age and cognitive development on task performance and of the role of 'fullness' in the comprehension of the word more. 3.31 Age Effects Table 8 presents percentage correct responses for M/L-l, M/L-2 and L/S-l for four age groups. A l l of the scores for the 2;6-2;11 group, except for less-2, are close to chance level. There is no constant in-crease in scores for the three older groups. The mean score for less-1 is considerably lower for the 3;6-3;ll group than for the age groups immediately older and younger. This is a reflection of the fact that two of the three subjects who showed a positive response bias on Task M/L-l and the subject who showed such a bias on half of the stimulus pairs were in this age group. Insert Table 8 about here. 55 TABLE S Experiments I and II. Correct responses (%) on Tasks M/L-l, M/L-2 and L/S-l for four age groups. Age Group 2;6-2;ll 3;0-3;5 3;6-3;ll 4;0-4;6 N 6 8 11 5 more-1 55.5 70.8 87.9 73.3 less-1 50.0 87.5 66.0 93.3 more-2 44.4 61.9 84.8 76.2 less-2 77.8 76.2 93.6 90.5 longer-1 50.0 70.8 90.9 73.3 shorter-1 61.1 87.5 93.9 80.0 56 3.32 Stimulus Effects Table 9 presents number of and percentage correct responses of 30 subjects for each of the three pairs of stimuli which were presented to a l l subjects. The total number of responses for Task M/L-2 is 29, since one child gave no responses to the questions. There are no significant differences in the number of times each stimulus was correctly chosen in Task M/L-l. For Task M/L-2, however, the greater quantity of Pair 3 was correctly labelled as more considerably less often than those of Pair 1 or Pair 2. The stimulus effect for the production of more is s t a t i s t i c a l l y significant (X 2 = 6.30, df = 2, p < .05). A division of the scores into age groups as presented in Table 10 shows that the lower number of correct productions of more (more-2) for Pair 3 is largely attributable to the youngest group: None of the 6 subjects in this group correctly labelled the row with more in the 5x3 comparison. This suggests that the use of 'fullness' as a criterion for the application of the word more tends to decline after the age of 3 years. The absence of stimulus effects in the data of Experiment II alone is explained by the fact that only one subject in that experiment was under the age of 3;0. Insert Tables 9 to 11 about here. TABLE 9 Experiments I and II. Correct responses (%) to Type A stimuli on Tasks M/L-l and M/L-2. Pair Stimulus Task M/L-l Task M/L-2 Percent N = 30 Percent N = 29 1 more (7) 73.3 75.9 less (5) 66.7 86.2 2 more (7) 76.7 82.8 less (3) 76.7 82.8 3 more (5) 70.0 51.7 less (3) 66.7 86.2 TABLE 10 Experiments I and II. Correct productions (%) of more for three stimulus pairs in Task M/L-2 for four age groups. Age Group 2;6-2;ll 3;0-3;5 3 ;6-3;ll 4;0-4;6 N 6 7 11 5 Pair 1 (7-5) 66.7 71.3 90.9 100.0 Pair 2 (7-3) 66.7 85.7 81.8 80.0 Pair 3 (5-3) 0.0 85.7 100.0 80.0 58 .* TABLE 11 Experiments I and II. Mean correct scores on Task M/L-l for thirty subjects, categorized according to behavior on linear ordering task (maximum score = 3 for each term). Stage 0 T Ha l i b . . I l l N 16 5 5 3 1 more 2.06 2.40 2.60 2.33 3.00 less 1.81 3.00 2.00 3.00 3.00 3.33 Linear Ordering Most of the subjects (21 of 30) were classified as Stage 0 or Stage I. Table 11 gives more-1 and less-1 scores for subjects categorized according to behavior on the linear ordering task. No relationship is evident between the a b i l i t y to reconstruct linear order and to identify more and less. There is no significant difference in the proportion of subjects at each stage level who were successful on Task M/L-l; i.e., who obtained a score of 2 or more out of 3 on both more and less (x 2 = 1.87, df = 3, p > 0.50). CHAPTER 4 DISCUSSION 4.0 Introduction It. was expected on the basis of the results of previous research that a substantial number of children would consistently choose the more numerous of the two sets in the more/less identification task, regard-less of verbal instruction. This particular response pattern, however, proved to be the exceptional one. The majority of children either consistently chose the correct set or made choices which appeared to be random. 4.1 Discussion of Results 4.11 Experiment I The positive response bias observed by Donaldson and Balfour (1968) and Palermo (1973) did not appear among any of the subjects in Experiment I. The difference between the mean correct scores for the positive and negative terms in both the more/less and longer/shorter identification tasks was in favor of the negative term. This finding was in the opposite direction to that expected. The mean scores for Task M/L-l and M/L-2 were similar, and in both cases the mean correct scores for less were higher, although in neither case was the difference s t a t i s t i c a l l y significant. With a maximum score of 3, i t is d i f f i c u l t to set a pass/fail criterion for individual scores. When the criterion is 100%, correct, 8 60 children f a i l for more and 5 for less. The corresponding figures for Task L/S are 9 failures for longer and 6 for shorter. 4.12 Experiment II The scores for Task M/L-l were closer to those expected on the basis of the results of previous investigations, insofar as the mean score for more was significantly higher than the mean score for less. In this group, also, the positive response bias (i.e., the consistent choice of the more numerous set) appeared, although in only 3 of the 16 sub-jects. A fourth subject gave this pattern of responses for one of the two sets of stimuli used in Task M/L-l (Type B); this same child showed a negative response bias on Task L/S-l. The c r i t e r i a given by Palermo (1974) of 0-2 and 6-8 errors for success and failure were applied to the scores on Task M/L-l. Four of the 16 subjects in Experiment II made 6-8 errors on identifying the set with less, ten made 0-2 errors, and two scored between these extremes. On the corresponding 'more' question, no children made more than 5 errors, 13 made 0-2 errors, and three scores f e l l between the extremes. Three of the four children who made 6 or more errors on identi-fying less made 2 or fewer errors on identifying more. These three children then, had the response pattern which Palermo (1973, 1974) considered to be indicative of the treatment of less as synonymous with more. The almost identical mean scores for Type A and Type B stimuli on Tasks M/L-l and M/L-2 are taken here to indicate that these children 62 as a group are able to use the cues of relative lengths of rows and relative 'crowding' of elements with equal f a c i l i t y in comparing quan-t i t i e s . The mean scores for longer and shorter on Task L/S-l are identi-cal. The three children who favored the more numerous set in Task M/L-l . did not favor either member of the pair in Task L/S-l. The scores for Task M/L-2 do not exhibit the pattern of those of Task M/L-l. For this task, there is no significant difference between the scores for more and for less. Likewise on Task L/S-2, there is no significant difference between the two words, but as in Task M/L-2 there is a slight difference in favor of the negative term. A comparison between the results of Tasks M/L-l and M/L-2 for the three subjects who consistently chose the greater quantity in Task M/L-l has important implications with regard to the interpretation of such a response pattern. According to Palermo's pass/fail c r i t e r i a , two of these children passed on both more and less in Task M/L-2; the third failed on bt>th. A fourth subject, favoring the more numerous set for only the Type B stimuli on Task M/L-l, attained a passing score for less and a f a i l i n g score for more on Task M/L-2. For at least these four sub-jects, then, any conclusion concerning their comprehension of the. compara-tives more and less, drawn on the basis of only one of these tasks, would be highly dependent on which of the two tasks was considered. The inadequacy of a task such as Task M/L-l as the only means of assessment of these children's comprehension of more and less is evident. 63 i 4.13 Combined Results and Comparison of Experiments I and II The apparent difference in group response patterns between Experiment I and Experiment II may be resolved by reference to the ages of the subjects in each group. Of the eight children who obtained higher scores for more than for less on Task M/L-l, five are between the ages of 3;6 and 3;9. No subjects in Experiment I happened to be in this age range. The only age group which had substantially higher mean scores on the identification of more than of less was the 3;6-3;11 group (see Table 8). The scores for Task M/L-l did not show a constant increase with age. Mean scores for the 3;6-3;11 group were higher for more and lower for less than those of the age groups both immediately older and younger. Although conclusions based on these data must be tentative due to the small group sample size, in view of similar findings by other investigators (e.g., Weiner, 1974; Eilers e_t a l . , 1974), i t appears that the positive response bias is particularly prevalent among children in the age range of about 3;6-3;11. For the group as a whole, there is a significant difference among the three pairs of stimuli which were used for a l l subjects in the number of times the label more is applied to the appropriate set. The set with 'more' was correctly labelled least often for Pair 3, a 5x3 comparison. This was the only pair of the three in which the set having more did not occupy the entire length of the cardboard strip. The stimulus effect was greatest for the 2;6-2;ll age group. None of the children in this age range correctly labelled the set with five candies as more than the set with three, while the two sets containing the maximum number of 64 • • • * candies were each c o r r e c t l y l a b e l l e d by four of the c h i l d r e n . The absolute d i f f e r e n c e between the numbers was the same for P a i r 3 and P a i r 1, and P a i r 3 was intermediate i n the r a t i o of i t s more to less numerous sets. The differ e n c e among the three pairs, then, must be due to the use by some c h i l d r e n of ' f u l l n e s s 1 as a cue i n applying the word 'more', a tendency which is most prevalent among the youngest c h i l d r e n . No connection between a b i l i t y to reconstruct l i n e a r order and performance on the more/less i d e n t i f i c a t i o n task was demonstrated. Only one of the c h i l d r e n who were categorized as being i n stage l i b or above made any errors on the M/L-l task; but the small number of c h i l d r e n i n the two highest categories and the a b i l i t y of some of the c h i l d r e n who were c l a s s i f i e d as Stage 0 or Stage I to c o n s i s t e n t l y i d e n t i f y more and less c o r r e c t l y precludes attachment of s i g n i f i c a n c e to these scores. There is a l s o no apparent connection between the p o s i t i v e response bias and any p a r t i c u l a r stage i n the a b i l i t y to reconstruct l i n e a r order. Of the three c h i l d r e n who demonstrated such a bias, two were at Stage 0 and one was at Stage I l a . 4.2 Comparison of the Present Results with those of Previous  Investigations The scores of subjects i n Experiment II on the i d e n t i f i c a t i o n of more and less (Task M/L-l) were c l o s e r to those reported by previous investigators (e.g., Donaldson and Balfour, 1968; Palermo, 1973, 1974) than were the scores of subjects i n Experiment I, although neither group exhibited a strong response bias i n favor of the greater quantity. 65 Palermo (1973) reported that ten of his thirty-two 3- and 4-year-old subjects knew the concept of less, and that those who did not behaved as i f they believed less to mean more. Donaldson and Balfour found that thirteen out of fifteen 3.5- to 5-year-olds responded in this manner. In the present study, only three children of a total of 30 had the re-sponse pattern which Palermo (1973, 1974) considered to be indicative of the treatment of more and less as synonyms. This proportion is very small in comparison with those reported by Donaldson and Balfour (1968) and Palermo (1973). The scores for the 3;6-3;11 group are similar to those of Weiner (1974) for her 3;5-3;ll and 3;0-4;4 age groups. Weiner also reported that the highest number of subjects who showed a positive response bias were in the age range of 3;5-3;11. In the present study, two of the three children who consistently chose the greater number in Task M/L-l and the one child who did so for Type B stimuli were in this age range. The higher score for less-1 in the 3;0-3;5 age group, when com-pared with the 3;6-3;11 group, is in agreement with the suggestion of Eilers et. a_l. (1974) that children under the age of 3;6 use a different response strategy for tasks involving comparatives from that used by children over this age. These authors associated the change in response pattern with a change in the meaning attributed to the words; i.e., children below age 3;6 extend the meaning of negative terms to cover positive terms, while children over this age give both words the meaning of the positive word. Such a conclusion, however, is not warranted by their data nor supported by that of the present investigation. In fact, 66 a comparison of individual scores on the M/L-l and M/L-2 tasks suggests that the percentage correct responses obtained from children on tasks intended to assess their comprehension of more and less is highly depen-dent on the task, and thus i t is unlikely that asymmetrical response patterns on the identification task have a semantic basis, as elaborated upon below. The performance which would be predicted on Task M/L-2 for a child who consistently chose the more numerous set in Task M/L-l depends on the interpretation of this response pattern. E, Clark's (1973) and Palermo's (1973, 1974) hypothesis, that less is understood to be a synonym for more, would predict that the child would find the question incomprehensible and would be particularly puzzled when asked to consider the member of a pair which contained less objects. The question would be interpreted as 'Does this one have more candies or more candies?'. The most probable responses to such a question would be either failure to answer or 'yes' when the comparison set had more candies and 'no' when i t did not. None of the children answered in this manner. Only one subject refused to answer the questions in the M/L-2 task, and this child did not show an asymmetry in her scores for Task M/L-l, but gave random responses, indicating comprehension of neither word. H. Clark (1970) suggests that more and less are both understood to mean 'some' and the set with more elements is consistently chosen because i t is a better example of 'some'. Again, the question 'Does this one have more candies or less candies?' would be given a different inter-pretation by the child, viz., 'Does this one have some candies or some 67 \ candies?'. In this case, the expected response would be 'yes' at a l l times or refusal to answer the question. No child answered in the former manner and, as mentioned above, the. only child who refused to answer the questions did not show a response bias on Task M/L-l. The results of the present research cast doubt on the hypothesis of Harasym et a l . (1968), that the differentiation of the meanings of more and less is preceded by the development of conservation. Thirteen of the 30 children in the present study had scores of 80% or better on a l l more/less tasks. Since the maximum age of the subjects was 4;6, i t is reasonable to assume that the majority of the children were non-conservers, yet almost half of them gave convincing evidence that they could d i f f e r -entiate more and less. The observation of Inhelder e_t a_l. (1974), that children who had not reached the level of conservation of quantity were able to correctly carry out instructions which required them to dis-tinguish more and less, further supports the possibility that children acquire the meanings of more and less before they develop the a b i l i t y to conserve quantity. Weiner's (1974) suggestion that children understand less correctly before they demonstrate perfect performance on a more/less identification task is supported by a comparison of the scores of some of the subjects on the two more/less tasks. Two of the three subjects who showed a marked preference for the more numerous set in the i d e n t i f i -cation task consistently and correctly labelled one member of a pair as more or less than the other; the third child appeared to choose the words more and less at random in Task M/L-2. 68 The possibility can most lik e l y be ruled out that knowledge of one of the words would be sufficient to enable the child to consistently choose the correct word in Task M/L-2, since i t is not at a l l clear that such a process of elimination lies within the intellectual capabilities of a 3-year-old child; i f such logic had been available to the child, then i t would ostensibly have been applied to Task M/L-l. It appears, then, that the phenomenon is artifactual, as Holland and Palermo (1975) suggest, but no explanation for why this particular group would respond in this manner suggests i t s e l f . The cognitive basis proposed for the response pattern observed in tasks of the type represented by Tasks M/L-l and L/S-l finds no support, due to the absence of any apparent relationship between the a b i l i t y to reconstruct linear order and performance on the more/less stimulus-choice task. It was not suggested that a causal relationship between the a b i l i t y to reconstruct linear order and the attainment of the concepts more and less might exist, but that a common i r r e v e r s i b i l i t y of thought might underlie both the d i f f i c u l t y in coordinating spatial relationships and in coordinating the reciprocal relationships of an asymmetrical pair. It appears, however, that some children are able to handle the compara-tive relationship before they are able to copy linear order. The only connection between the two a b i l i t i e s is that they both seem to develop between the ages of 3 to 5 years,. Although Holland and Palermo's (1975) suggestion that the ' less-is-more phenonemon' may be an experimental artifa c t is supported by the present study, the observation that some children favored the negative 69 » member of the pair in the more/less and longer/shorter identification tasks does not confirm their proposal that the child invariably sees the greater quantity as the better choice, regardless of verbal instructions. As explanation for the bias in terms of the perceptual salience of longer rows (as suggested by Weiner, 1974) is denied by several factors. For example, the three subjects who consistently chose the longer row cf objects did riot show any bias in favor of the longer stimulus'--in identi-fying longer and shorter. Furthermore, the bias in favor of the more numerous set was as strong when the paired comparisons were equal in length as when the row with more was longer. There was also one i n c i -dence of a reversal from a positive response bias on Task M/L-1(B) to a negative bias on the longer/shorter identification task. The preoccupation in the past with the apparent confusion of less with more in young children has constituted a diversion from progress toward an understanding of how such concepts as more and less develop. Piaget (1954) suggests that the earliest conceptions of comparative words are of absolute rather than relative qualities. Values such as 'more' or 'less' are attributed on the basis of subjective judgments. The greater tendency for children to use the word more for a set in which a l l the. available spaces are occupied indicates that the word more may not at f i r s t be used in the comparative sense, but rather on the basis of the absolute characteristic of 'fullness'. Weiner (1974) and Donaldson and McGarrigle (1975) reported a tendency on the part of very young children to use 'fullness' as a cue in identifying the member of a pair with more. Weiner observed that her youngest subjects more often 70 .* correctly identified the row with more objects when the space was f i l l e d , and Donaldson and McGarrigle found that some children incorrectly iden-t i f i e d a set in which a l l the spaces were f i l l e d as more than a longer, more numerous set to which empty spaces were added. It is impossible to determine whether 'fullness' is a component of the meaning of the word more for young children or is an effect of the tendency of these children to make absolute rather than relative judgments of quantity. Indeed, i t is perhaps meaningless to consider the semantic components of a word as separable from the perceptual cues used in identifying its referents. 4.3 Summary and Conclusions For some children, two types of task, both of which have been considered to be a definitive assessment of children's comprehension of more and less, yield different response patterns and i f considered separately would lead to different conclusions regarding the understand-ing of these terms. The evidence presented here strongly favors a non-semantic (i.e., non-Unguis tic) basis for the positive response bias shown by some c h i l -dren on stimulus-choice tasks involving comparatives. As a consequence, drastic revision of the 'Semantic Feature Hypothesis' is indicated, since the data relating to children's treatment of quantitative and dimensional comparative adjectives form a major part of its empirical foundation. Further examination of the performance of children on tasks of the nature of Task M/L-l does not promise to offer enlightenment with regard to their understanding of the words in question. The importance has been demon-strated of using extreme caution in interpreting children's behavior under a r t i f i c i a l conditions in terms of their comprehension of the words in-volved. There is no indication of a relationship between mobility of thoughts as measured by the a b i l i t y to coordinate relationships of proximity and the capacity to deal with problems involving comparisons. The suggestion that 'fullness' is part of the earliest conception of more is supported. This is interpreted as a reflection of the tendency of young children to make judgments in terms of absolute rather than relative c r i t e r i a . 72 REFERENCES Amidon, A., and Carey, P. 1972. Why five-year-olds cannot understand before and after. Journal of Verbal Learning and Verbal Be-havior 11. 417-423. Beilin, H. 1968. Cognitive capacities of young children: a replication. Science 162. 920-921. Bever, T.G., Mehler, J., and Epstein, J. 1968. What children do in spite of what they know. Science 162. 921-924. Bloom, L. 1970. Language Development: Form and Function in Emerging  Grammars. Cambridge: The M.I.T. Press. Campbell, R.N., and Wales, R.J. 1969. Comparative structures in English. Journal of Linguistics 5_. 215-251. Clark,- E.V. 1971. On the acquisition of the meaning of before and after. Journal of Verbal Learning and Verbal Behavior 10. 266-275. Clark, E.V. 1973. What's in a word? On the child's acquisition of semantics in his f i r s t language. In Moore, T.E. (Ed.), Cognitive  Development and the Acquisition of Language. New York; John Wiley and Sons. 65-110. Clark, H.H. 1970. The primitive nature of children's relational con-cepts. In Hayes, J.R. (Ed.), Cognition ana the Development of  Language. New York: John Wiley and Sons. 269-278. Donaldson, M., and Balfour, G. 1968. Less is more: a study of language comprehension in children.. British Journal of Psychology 59. 461-471. Donaldson, M., and McGarrigle, J. 1974. Some clues to the nature of semantic development. Journal of Child Language 1_. 185-194. Donaldson, M., and Wales, R.J. 1970. On the acquisition of some relational terms. In Hayes, J.R. (Ed.), Cognition and the  Development of Language. New York; John Wiley and Sons. 235-268. Eilers, R.E., Oiler, D.K., and Ellington, J. 1974. The acquisition of word meaning for dimensional adjectives: the long and short of i t . Journal of Child Language I. 195-204. Ferreiro, E., and Sinclair, H. 1971. Temporal relationships in language. International Journal of Psychology 6. 39-47. Griffiths, J.A., Shantz, C.A.,.and Sigel, I.E. 1967. A methodological problem in conservation studies: the use of relational terms. Child Development 38. 841-848. 73 Harasym, C.R., Boersma, F.J., and Maguire, T.O. 1971. Semantic differential analysis of relational terms used in conservation. Child Development 42. 767-779. Holland, V.M., and Palermo, D.S. 1975= On learning 'less': language and cognitive development. . Child Development 46. 437-443. Inhelder, B., Sinclair, H., and Bovet, M. 1974. Learning and the Development of Cognition. Cambridge: Harvard University Press. Klatsky, R.L., Clark, E.V., and Macken, M. 1973. Asymmetries in the acquisition of polar adjectives: linguistic or conceptual? Journal of Experimental Child Psychology 16. 32-46. Mehler, J., and Bever, T.G. 1967. Cognitive capacity of very young children. Science 158. 141-142. Palermo, D.S. 1973. More about less: A study of language comprehension. Journal of Verbal Learning and Verbal Behavior 12. 211-221. Palermo, D.S. 1974. S t i l l more about the comprehension of less. Developmental Psychology 10. 827-829. Piaget, J. 1952. The Child's Conception of Number. London: Routledge & Kegan Paul. Piaget, J. 1968. Quantification, conservation and nativism. Science  162. 976-979. Piaget, J. 1969. The Child's Conception of Time. London: Routledge & Kegan Paul. Piaget, J., Grize, J.B., Szeminska, A., and Bang, V. 1968. Epistemologie  et psychologie de la fonction. Paris: Presses Universitaires de France. Cited in Ferreiro and Sinclair, 1971. Piaget, J., and Inhelder, B. 1956. The Child's Conception of Space. New York: W.W. Norton. Piaget, J., and Inhelder, B. 1964. The Early Growth of Logic in the  Child; Classification and Seriation. New York; Harper and Row. Piaget, J., Inhelder, B., and Szeminska, A. 1960. The Child's  Conception of Geometry. New York; Basic Books. Sinclair, H. 1969. Developmental psycholinguistics. In Elkind, D., and Flavell, J.H. (Eds.), Studies in Cognitive Development. New York: Oxford University Press. 315-336. Weiner, S.L. 1974. On the development of more and less. Journal of  Experimental Child Psychology 17. 271-287. 

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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0100152/manifest

Comment

Related Items