FACTORS INFLUENCING THE DEVELOPMENT for ' . PARADIGMATIC AND SYNTAGMATIC ASSOCIATIONS. IN ‘ ' SIMPLE ARTIFICIAL LANGUAGES ‘ - Thesis for the Degree of PII. D. MICHIGAN STATE UNIVERSITY WANDA R. .IAGOCKI 1969 LIBRA R Y 1 MlCl’llguI: Late 5 Univc—rsfi 1 LI This is to certify that the thesis entitled Factors Influencing the Development of Paradigmatic and Syntagmatic Associations in Simple Artificial Languages presented by Wanda R. Jagocki has been accepted towards fulfillment of the requirements for _Eh.._D_.__ degree in _Ea¥chnlngy ."I 1") i ' . 9 ° / /“" " ’I ‘l' I f’ k’ “('1 ‘r' "X // .A "- ' S. r r“ I Major professor Da‘e May 16, 1969 0-169 w M4? ABSTRACT FACTORS INFLUENCING THE DEVELOPMENT OF PARADIGMATIC AND SYNTAGMATIC ASSOCIATIONS IN SIMPLE ARTIFICIAL LANGUAGES By Wanda R. Jagocki The present study contrasted two hypotheses proposed to account for the development of paradigmatic associa- tions in natural language. One hypothesis, the inter- substitution hypothesis, suggests that paradigmatic associ- ations result from the use of words by §s in identical speech contexts. The intersubstitution hypothesis has been advanced by Deese, Horowitz, and McNeill, among others. A second hypothesis, suggested by Ervin (1961) and others, maintains that paradigmatic responses may re- sult because of the simultaneous elicitation of several competing responses. This hypothesis has in its favor the dual explanation of both paradigmatic associations and so-called "Spoonerisms." To investigate these two hypotheses in an explora- tory way, 32 §s were presented with nonsense figures that varied in size as well as form. These nonsense figures were assigned nonsense syllable labels which could refer to either form (noun) or size (adjective). A discrete word association test in which nonsense syllable elements Wanda R. Jagocki were presented as stimuli was administered after §s reached a criterion of learning, and again after they reached a more stringent criterion. Both the associative response and its latency were recorded. Associative re- sponses could be categorized not only as syntagmatic or paradigmatic, but also as contiguous, positional, or non- positional. The major conclusions were as follows: 1. Subjects were able to learn artificial languages with relative ease, i.e., within a one hour session. 2. Classification of artificial language word asso- ciation data in terms of grammatical class indicated that the majority of the associations were syntagmatic, and could be attributed to the contiguous appearance of the associates in artificial language sequences or contexts. 3. Two findings were taken as evidence that re- sponse interference had been generated as a result of artificial language training. First, the proportion of contiguous noun-adjective (N-Ac) word associations sig- nificantly exceeded (p < .05) the proportion of contiguous adjective-noun (A-Nc) associations; and the second, median A-Nc associative latencies were significantly slower (p < .05) than the median N-Ac associative latencies. A. Under the conditions of the present study para- digmatic adjective associations could develOp only as a means of resolving interference while certain paradigmatic Wanda R. Jagocki noun associations could develop as a result of intersub— stitution. Statistical analysis indicated that within the framework of the artificial languages used, there was evidence (p < .05) that intersubstitution was involved in the development of paradigmatic noun associations. 5. There was evidence that interference and appear- ance in identical positions within different contexts did not interact in an additive manner to produce associa- tions. 6. Associative symmetry of contiguous syntagmatic associations as measured by associative latencies was observed. The relationship between paradigmatic association develOpment and other variables such as ability to verbal— ize artificial language rules, amount of foreign language training, frequency of paradigmatic associations to English words, and type of artificial language training is discussed. FACTORS INFLUENCING THE DEVELOPMENT OF PARADIGMATIC AND SYNTAGMATIC ASSOCIATIONS IN SIMPLE ARTIFICIAL LANGUAGES By . "X .. (.. ‘-.5.,\“ " Wanda RLLJagocki A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1969 ACKNOWLEDGMENTS 1I gratefully acknowledge the assistance of Dr. William T. Stellwagen in the formulation and interpre— tation of this study. I wish to thank the other members of my committee, Drs. A. Barch, M. R. Denny, G. Hatton, and S. C. Ratner for their contributions to this pro- ject. I also wish to thank Stanley Cohen and Cynthia Haas for their invaluable help, and to acknowledge my two assistants, Janet Blomberg and Patsy Speer. ii TABLE OF CONTENTS ACKNOWLEDGMENTS LIST OF TABLES . . . . LIST OF FIGURES Chapter I. INTRODUCTION Dependent Variables in the Word Association Studies . . . . Literature Related to Form Class of Word Associations. . . Syntagmatic Association . Paradigmatic Association . . . Esper Studies. . . . . . Pilot Work . . A Simple Artificial Language to Com- pare the Effects of Interference and Intersubstitution on Paradigmatic Association Development. . . Major Hypotheses. II. EXPERIMENT I Method . Results. . . Discussion. . . . III. EXPERIMENT II . . . . Method . . . . . . . Procedure . . . . . . . . IV. RESULTS. . . . . . . . .Artificial Language Word Association T681: 0 0 O O O O O O 0 Observed Proportions Tested Against Theoretical Proportions. iii Page ii vii l\) HA4 n+4U12xo 15 I7 21 21 23 23 28 28 29 43 “3 48 Chapter Evidence of Interference . . . Interference vs. Intersubstitution. Evidence of Intersubstitution . . Evidence of Interaction of Factors. Artificial Language Training. V. DISCUSSION. . . . . . . . . . . Interference vs. Intersubstitution in Paradigmatic Association Development Evidence of Interference . . . Evidence of Intersubstitution Evidence of Interaction of Factors. Contiguous Syntagmatic Associations Associative Direction of Contiguous Response. . . . . . . Suggested Researc . . . : VI. CONCLUSIONS . . . . . . . . . REFERENCES. APPENDICES. . iv Page 52 62 6A 68 72 73 76 77 80 81 82 89 91 95 LIST OF TABLES Table Page 1. A simple artificial language . . . . . l3 2. Paradigm of a simple artificial language in which adjective elements occupy different contextual positions . . . . . . . l6 3. Frequency and percentage of associates to six nonsense syllable stimuli on pre- sentations l and 2 of the syllable questionnaire. . . . . . . . . . 24 A. Values of x2 obtained for the frequency distribution of responses to each of six nonsense syllables on presentations I and 2. . . . . . . . . . . . 25 5. Rank order of nonsense syllable associa- tions according to mean percent on pres- entations l and 2 . . . . . . . . 26 6. High and low adjective association-- artificial languages . . . . . . . 27 7. Classification of associations made in artificial languages . . . . . . . 45 8. Observed proportions and average median latencies in seconds of associations of Artificial Language Word Association Tests for High and Low Artificial Language Groups . . . . . . . . . A9 9. Summary of z—tests for observed proportions of associative classes vs. theoretical proportions . . . . . . . . . . 51 10. Summary of z-tests for difference between observed proportion of A-Nc and N-Ac Associations on Word Association Tests 1 and 2 for High and Low Groups. . . . 54 Table 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. Average median latencies in seconds of forward and backward contiguous syntagmatic associations for High and Low Groups on Tests 1 and 2 Analysis of variance of median latencies of forward and backward contiguous syntagmatic associations for High and Low Groups . . . . . . . . . . . . Analysis of variance of median latencies of A— No and N- Ac associations for High and Low Groups. . . . . . . . Summary of z-tests for difference between observed proportion of A-A and N-Np associations for High and Low Groups Summary of z-tests for difference between observed proportion of N-Np and N-an associations for High and Low Groups Summary of z—tests for difference between observed proportion of A-Np and N—Ap associations for High and Low Groups Summary of z-tests for difference between observed proportion of A-Np and A-A associations for High and Low Groups Average median latencies in seconds and standard deviations of associations to English Word Association Tests 1 and 2 and Artificial Language Association Tests 1 and 2.. . . . Analysis of variance for median associative latencies on English and Artificial Language Association Tests 1 and 2. Simple artificial languages in which the same stimulus shapes appear in different sizes . . . . . . . . . . Simple artificial language in which noun elements occupy different contextual positions . . . . . . . . vi Page 55 56 57 62 63 65 67 71 71 8A 87 LIST OF FIGURES Figure Page 1. "Small" Size Nonsense Shape Stimuli . . . 30 2. "Middle" Size Nonsense Shape Stimuli. . . 31 3. "Large" Size Nonsense Shape Stimuli . . . 32 vii CHAPTER I INTRODUCTION Empirical approaches to the study of word associa— tion began with the collection of normative data. Even a casual inspection of the norms indicates that stimulus words and response words are frequently of the same gram- matical class. This observation raises questions since a frequency view of learning, with or without reinforcement, has become more widely accepted. With respect to natural language, the development of paradigmatic associations, i.e., associations between stimuli and responses of the same grammatical class, has generally been attributed to the use of words in identical speech contexts. This hypothesis has been called the intersubstitution hypothe— sis by McNeill (1963, 1966) and others. According to this View, words which appear in the same position within a sentence frame will become associates of each other. Another view of the association process would suggest that in the past a number of response words have been rein— forced in the presence of a given stimulus word. Thus, considerable response competition would result if the given stimulus word were presented without sufficient contextual isolation. In the present study, an artificial language situa— tion, rather than a natural language setting, was employed to examine the factors operating in the deveIOpment of paradigmatic associations. The Ss were given a word asso- ciation test in which elements of a previously learned artificial language were presented as stimuli. The major purpose of the present study was to contrast the inter- substitution and interference hypotheses of paradigmatic association development. Dependent Variables in the Word Association Studies Before proceeding further, mention of the dependent variables involved in a discrete word association test should be made. In a discrete word association situation a subject is typically instructed to respond with the first word that occurs to him when he sees or hears a stimulus word. Response latency or reaction time, and commonality are generally used as indices of the strength of the associa- tion between the stimulus and response. Commonality re- fers to the frequency of the associative response in a normative sample. It is assumed that the stronger the associative strength the shorter the associative latency and the greater the commonality. Schlosberg and Heiniman (1950) found a high (-0.80) correlation between the two associative strength indices. In a recent review of English word association studies Jung (1966) pointed out "recently reaction time has been virtually replaced by commonality as the most widely used measure . . ." (p. 129). The commonality measure has been used almost ex— clusively in the current literature related to form class of word associations. Literature Related to Form Class of Word Associations The most common method of categorizing associations in terms of form class or grammatical class is patterned after Fries (1952). "In this analysis, all words which can occupy equivalent positions within English utter- ances are declared to be members of the same grammatical class" (Deese, 1965, p. 99). When word associations are classified according to such an analysis, it is observed that the form class of the associative response tends to vary as a function of the grammatical class of the stimu— lus word. The word associations of young children (under l the age of six) tend to be "syntagmatic,' i.e., responses are of a different form class than the stimuli. The asso- ciations of older children and adults are "paradigmatic," i.e., responses are of the same form class as the stimuli (Brown and Berko, 1960; Entwisle, Forsyth and Muuss, 196“, Ervin, 1961; Fillenbaum and Jones, 1965). However, Deese (l962a)has qualified the syntagmatic—paradigmatic find- ings as follows: The older generalization that adult associa— tions are largely paradigmatic is unconditionally true only for nouns. Adjectives and verbs are about equally syntagmatic and paradigmatic; ad— verbs yield largely syntagmatic associations (p. 81). Syntagmatic Association "Syntagmatic associations are, in general, sequen- tial elements or at least elements which usually occupy different positions within phrases or sentences" (Deese, 1965, p. 103). Thus, syntagmatic responses are generally contiguous with their stimuli in language sequences. "As in paired-associate learning, it is assumed that one fac— tor critical to the formation of associative bonds is the experience of words in contiguity" (McNeill, 1966, p. 5A8). McNeill also suggested that the discrete word association test may be likened to the recall test in a paired- associate learning situation. If such a comparison can be made, certain observations related to paired—associate learning may also have some bearing on syntagmatic asso— ciation development. In paired associate learning situations recall in the forward direction has generally been found superior to recall in the backward direction (Ekstrand, 1966). However, Asch and Ebenholtz (1963) proposed that "when an association is formed between two distinct terms, a and b, it is established simultaneously and with equal strength between 6 and a." This "principle of associative symmetry" is hypothesized to operate under conditions in which both the stimulus and response elements are equally "available." Availability has been defined as an item's "accessibility to recall," or the proportion of §S who recall an item correctly (Horowitz, Norman and Day, 1966). Horowitz, Brown and Weissbluth (1964) had subjects learn pairs of associates in which the "stimuli of some pairs were responses in other pairs; these stimuli became avail- able during original learning" (p. 5A2). A free associa- tion test following traning indicated that "a backward association occurred as readily as a learned forward association if the PA stimulus was available" (Horowitz, §t_ag,, 196A, p. 5A1). In summary, the development of associations between stimuli and responses which are of different form classes is generally attributed to the contiguous appearance of the associates in language sequences. There are data which indicate that syntagmatic associations do not necessarily develop only in a forward direction. Avail- ability and associative symmetry may also influence syn- tagmatic association. Paradigmatic Association Traditional association theory has emphasized contiguity as the most potent factor operating in the production of word associations. Ordinarily words of the same form class do not appear in contiguity within sentences. Thus, it is difficult to posit contiguity as a variable in the development of paradigmatic association. However, Ervin (1961) hypothesized that words of the same form class could be placed into contiguity when a listener, trying to anticipate what he will hear next, makes an "erroneous anticipation." For example, if one hears "a cup of . . . ," and anticipates "coffee," but hears "tea," then the two nouns are placed in contiguity. McNeill (1966) tested Ervin's "erroneous anticipation" hypothesis by presenting subjects: pairs of nonsense syllables in sets of English sentence frames, with each pair appearing in a separate set. The members of pairs, sub- stituting for one another equally often, formed idealized "grammatical classes" — that is, pairs of words that enjoyed identical privileges of occurrance . . . Ss were required to make overt anticipations of nonsense syllables upon presen- tation of sentence frames, a procedure that brings the process of erroneous anticipation to the sur- face (p. 549). According to Ervin's hypothesis an increase in the fre- quency of anticipations should lead to an increase in the frequency of paradigmatic responses. McNeill's results did not favor the hypothesis that paradigmatic associa— tions result from contiguity of erroneous anticipations of speech. The most popular hypothesis concerning the develop- ment of paradigmatic associations is one which has been referred to as the "intersubstitution” hypothesis (Deese, 1962b, 1965; Ervin, 1961; Horowitz, et a1., 1963; Horowitz, Norman and Day, 1966; McNeill, 1963). Accord— ing to the intersubstitution notion, "paradigmatic asso— ciation results from the use of words in identical speech contexts” (McNeill, 1963, p. 250). To investigate the intersubstitution hypothesis, McNeill (1963) again used a procedure in which nonsense syllables appeared in English sentence frames. Certain syllables appeared in adjective positions while others appeared in noun positions within the sentences. The §S were given either 20, MO or 60 training trials. Train— ing was followed by a free association test in which noun syllables were stimuli, a recall test in which adjective syllables were stimuli, and a usage test in which §S were asked to make up sentences with the syl— lables. McNeill (1963) found that: the frequency of association between non- sense syllables presented as nouns in identical English contexts increased as a function of the number of presentations (p. 259). However, there was no correlation between the frequency of paradigmatic association and the probability of using the artifical words in the same grammati- cal class as was imposed on them in training (p. 262). Braine (1963; 1965) offered a version of the inter— substitution hypothesis and prOposed that grammatical structure is acquired by "contextual generalization." For verbal learning, contextual generalization may be defined informally as follows: when a subject, who has experienced sentences in which a segment (morpheme, word or phrase) occurs in a certain position and context, later tends to place this segment in the same position in other contexts, the context of the segment will be said to have generalized, and the subject to have shown contextual generalization (1963, p. 323). Thus, ”'what is learned' are primarily the proper loca— tions of words in sentences" (Braine, 1963, p. 32A). Braine demonstrated contextual generalization by having children learn "miniature artificial languages with non- sense syllables as words" (Braine, 1963, p. 324). In one of Braine's "languages" "there were two classes of words, A words and P words, and sentences were always two words long and consisted of an A word followed by a P word" (Braine, 1963, p. 325). Language training con- sisted of sentence—completion problems in which two A words and two P words were used. Following initial learning, the Ss were given generalization problems in which a new A or P word was presented, and SS were to supply an appropriate learned alternative word. On the basis of the results of the generalization test, Braine concluded that: subjects who have experienced sentences in which words occur in a certain position and con- text tend to place these words in the same posi- tions in new contexts. Such behavior indicates the learning of an association of words with their positions, the context generalizing (1963, p. 326). Similar results were also obtained when more elaborate "languages" were used. A theoretical explanation of paradigmatic associa- tion development which embodies the competing response notion has been largely ignored by most investigators in the area, with the exception of Ervin (1961). The com- peting response idea and its Operation in the word asso- ciation test situation has been described as follows: The stimulus word, through past experience, has become associated with many different re— sponse words and is capable of eliciting any of them. When the stimulus word is presented, the potential responses compete among them- selves, and the strongest connection wins and governs the overt response. The frequency and speed with which a response wins out vary directly with the response's own strength, and inversely with the strength of its competitors (Woodworth and Schlosberg, 1954, p. 49). The competing response or interference view can be applied as an explanation of paradigmatic association development. Assuming that a stimulus word is capable of eliciting several, equally strong, competing syntagmatic responses, a subject who is instructed to respond to the stimulus word as quickly as possible must resolve the existing interference. The alternatives available to the subject are: (l) to respond with a comparatively long latency with one of the competing responses, i.e., to give a syntagmatic response; (2) to respond with a Spoonerism, or a response which contains parts of two or more of the competing responses; or (3) to respond with a word out- side of the competing response hierarchy. Such a word might be of the same form class as the stimulus word. In the present study, interest is focused on the development of paradigmatic associations of nouns and lO adjectives. All hypotheses and interpretations stated from this point on are with respect to these two form classes only. Deese (1965) noted that "nouns are para— digmatic whether they are common nouns or rare nouns common adjectives are somewhat more likely to be paradigmatic than are uncommon adjectives" (p. 106). The interference View would account for the finding that common adjectives tend to elicit paradigmatic responses by arguing that common adjectives would elicit many equally strong competing responses, while uncommon adjec- tives would elicit weak response competition. Thus the interference elicited by a common adjective would be re- solved by responding paradigmatically. If an uncommon adjective elicits little or no interference, the strong- est contiguous response or a syntagmatic association could be expected. The interference hypothesis would account for the observed paradigmatic responses to nouns by main— taining that through past experience a noun would elicit several associative responses. Responding paradigmatically with a word which is outside of the competing response hierarchy could resolve the existing interference. In any study of natural language behavior, individual differences in previous verbal experiences represent a potent uncontrolled variable. A simple artificial language situation, such as that used by Braine (1963), would allow investigation of word association development with 11 previous experience under experimental control. The use of a simple artificial language stems from early studies done by Esper (1933). Esper Studies Esper (1933) held that "language, in its fundamental, living form, consists of a system of verbal responses to (chiefly external) stimulus patterns . . . " (p. 347). Thus, a simple artificial language was designed in which nonsense figures (stimuli) were assigned nonsense names (responses). The conditions as described by Esper (1933) were as follows: (a) verbal responses were attached to a number of stimulus—objects which resembled one another in shape or size, but which were so selected that there was a certain asymmetry in the objective classes to which they belonged; and (b) addi- tional objects were subsequently introduced which systematically resembled the original objects but for which no specific names had been learned. In both cases, the purpose was to provide condi- tions favoring varying degrees of instability in the stimulus-response relationship (p. 347). Subjects were trained under various conditions for a total of eight consecutive months. One of the major conclusions made with regard to the associative process was that: Stimuli resulting from one verbal response be— come capable of eliciting a specific other verbal response whenever there is a frequent recurrence of environmental situations which tend to elicit both responses simultaneously or successively. An environmental situation might elicit such multiple responses either because it contains both of the objects a and g each of which tends to elicit its own specific response, or because it contains an object a which tends to elicit 12 not only its own specific response but also the response specific to another (absent) object I; (p. 376). Esper's approach and findings have been virtually ignored in subsequent literature on the development of association. One probably reason for this is the long training period of the subjects. However, the use of simple artificial languages would allow experimental in- vestigation of paradigmatic association development. The following rationale can be used to generate simple arti- ficial languages. A single visual form can be verbally identified by a learned label, and such a label or shape name may be classed as a noun. If the visual form varied along a dimension such as size, a verbal unit may be assigned to each value of the size dimension. The verbal units used to signify size may be classed as adjectives. Thus, each instance of a visual form could be uniquely identified with two verbal units, one referring to form size and the other referring to form name. For example, a simple artificial language could be generated as shown in Table 1. As in Esper's design, the stimuli are nonsense shapes. Each shape is associated with two pronounceable nonsense syllables. One of the syllables functions as an adjective, while the other syllable functions as a noun. In the arfitifial language in Table l, the syllable "ged" appears as a part of the response only when either 13 TABLE l.—-A simple artificial language. Learned Response .. Stimulus Size Nonsense Shape Adjective Noun Size I A1 ged faw Size 2 A2 hib faw Size 1 Bl ged mep Size 2 B2 hib mep nonsense shape of Size 1 is presented, while the syllable ”hib" appears only when either nonsense shape of size 2 is presented. Therefore, "ged" and "hib" serve as ad— jectives, while "faw" and "mep" serve as nouns. The adjective syllable "ged" appears in contiguity with the noun syllable ”faw" and the noun syllable "mep." As training progresses, the syllable "ged" should be asso— ciated to an equal degree with both noun syllables, "few" and "mep." Similarly, the adjective syllable "hib," which has appeared in contiguity with the noun syllable "faw" and the noun syllable "mep," should elicit both noun syllables with equal strength at the end of train- ing. Assuming that backward associations do develop, each of the noun syllables should be associated to the same degree with each of the adjective syllables. Thus, once such a simple artificial language has been learned, it is hypothesized that a situation exists which is 14 similar to that cited previously by Esper (1933, p. 376). In other words, paradigmatic associations would be ex- pected to develop since the stimuli for each individual response element elicits two simultaneous or successive syntagmatic associates with approximately equal strength. Pilot Work Pilot studies indicated that SS are able to learn simple artificial languages, such as that presented in Table 1, within an hour session. After learning the simple artificial languages to a predetermined criterion, Ss were given a discrete free association test in which each of the artificial language nouns and adjectives were presented as stimuli and the SS were instructed to re- spond with the first nonsense word, of those that they had just learned, that they thought of. The associative responses and their latencies were recorded. Classifica- tion of the associative responses as paradigmatic (i.e., noun-noun or adjective—adjective) or syntagmatic (i.e., adjective-noun or noun-adjective) was made. It was found that although §s gave a preponderance of syntag- matic responses, with adjective—noun (forward associa- tions) occurring as frequently as noun-adjective (back- ward associations), some paradigmatic responses were also given. In the simple artificial languages used in pilot studies, the nonsense syllable adjectives always appeared 15 as the first element of the two syllable response to a nonsense shape, while noun syllables always appeared as the second element of the response. It could be argued that the development of any paradigmatic associations was based on intersubstitution, i.e., use in identical speech contexts, rather than associative competition or inter- ference. In order to control for the possibility that two grammatically similar elements might become asso- ciated because they appear in a similar position within verbal contexts, it is necessary to design a simple artificial language in which some elements are of the same form class but do not occupy identical positions, and some are of the same form class and occupy identical positions. A Simple Artificial Language to Compare the Effects of Interference and Inter- substitution on Paradigmatic Association Development In the rationale suggested earlier for generating simple artificial languages, an adjective was described as a verbal unit which signified a particular value along some dimension, such as size. Acceptance of such a definition enables one to assume that there are no limi- tations on the position an adjective might occupy within a verbal utterance. Therefore, a possible paradigm of a simple artificial language in which verbal elements of the same form class do not appear in similar contextual positions appears in Table 2. 16 TABLE 2.--Paradigm of a simple artificial language in which adjective elements occupy different contextual positions. Stimulus Response Size Nonsense Shape Nonsense Syllables Size 1 W1 Adjective l - Noun 1 Size 1 X1 Adjective l - Noun 2 Size 2 Y2 Noun 3 - Adjective 2 Size 2 Z2 Noun 4 - Adjective 2 In the paradigm presented in Table 2, each adjective syl- lable should be associated to an equal degree with two noun syllables. It is maintained that only the response interference hypothesis would predict the development of any association between the two adjectives, while the intersubstitution hypothesis would predict the development of associations between nouns appearing in identical con— texts. Therefore, in a word association situation in which artificial language elements are presented as stimuli, S might respond with any one of the five remaining syl- lables. With reference to Table 2, when an adjective syllable is presented as a stimulus S might respond with: (1) the other adjective element (Adjective—Adjective paradigmatic association); (2) one of the two noun syl- lables which appeared contiguously with the adjective 17 stimulus during training (Adjective—Noun contiguous syntagmatic association); or (3) one of the two noun syl- lables which had not appeared contiguously with the adjec- tive stimulus, but had occupied the same contextual posi— tion as the adjective (Adjective-Noun positional syntagmatic association). When a noun syllable is pre- sented as a stimulus in the word association test situa- tion the S might respond with: (l) the other noun syllable which occupied the same contextual position and had also appeared contiguously with the same adjective syllable during training (Noun-Noun positional paradigmatic asso— ciation); (2) one of the other two noun syllables which had not occupied the same contextual position as the stim- ulus noun (Noun-Noun non-positional paradigmatic asso- ciation); (3) the adjective syllable which had appeared contiguously with the stimulus noun during training (Noun- Adjective contiguous syntagmatic association); or (4) the adjective syllable which had not appeared contiguously with the noun stimulus, but which occupied the same con- textual position during training (Noun-Adjective posi- tional syntagmatic association). Major Hypotheses During Artificial Language Training each adjective syllable appears with equal frequency in contiguity with two noun syllables. At the end of training a word asso- ciation test is administered in which artificial language 18 elements are presented as stimuli and the associative responses and their latencies are recorded. The appear- ance of a noun artificial language syllable in a word association test should elicit only one contiguous adjec- tive syllable and response competition or interference. However, the appearance of an adjective syllable as a word association stimulus should elicit two equally strong contiguous noun associates and response interference. Evidence of Interference If response interference is generated as a result of training, the following word association test results are expected: 1. The prOportion of noun—adjective contiguous syntagmatic associations should exceed the proportion of adjective-noun contiguous syntagmatic association. In other words, if interference is generated by the presen- tation of an adjective stimulus, rather than responding with one of the competing contiguous noun responses, §S would be more likely to resolve the interference by re- sponding in some alternative fashion. 2. The average median latency of adjective-noun contiguous syntagmatic associations should exceed the average median latency of noun-adjective contiguous syn- tagmatic association. That is, if an adjective stimulus elicits response interference then §S should respond more slowly to an adjective than to a noun. l9 Interference vs. Intersubstitution If interference rather than intersubstitution is the major factor involved in the deveIOpment of paradig- matic association, the proportion of adjective-adjective associations should exceed the porportion of noun-noun positional associations. It is maintained that adjective- adjective associations would develop as a result of inter- ference, while noun-noun positional associations would develop as a result of intersubstitution. Evidence of Intersubstitution If the development of paradigmatic associations is a result of intersubstitution, i.e., association as a re- sult of appearance in the same context, then the prOpor- tion of noun-noun positional associations should exceed the proportion of noun-noun non—positional associations. Noun—noun non-positional associations would presumably be the result of chance. Evidence of Interaction of Factors Earlier it was stated that one manner in which inter— ference might be resolved is by responding with an element which is outside of the competing response hierarchy. In the present study both adjective-adjective and adjective- noun positional associations illustrate this mode of re- sponding. However, it is possible that associations develop between elements which have appeared in identical 20 positions but within different contexts (Braine, 1963). An adjective-noun positional association is an example of such an association. Therefore, an adjective-noun posi— tional association could be a result of interference, or a result of mediation in terms of position, or both. While an adjective—adjective association could develop only as a result of interference, a noun-adjective asso- ciation, on the other hand, could develOp only as a re— sult of appearance in identical positions within different contexts. If the effects of interference and position interact in an additive manner, then it would be pre- dicted that: (1) the prOportion of adjective—noun posi— tional associations should exceed the proportion of noun- adjective positional associations, and; (2) the proportion of adjective-noun positional associations should exceed the porportion of adjective-adjective associations. The present study consists of two experiments. The results of Experiment I are used in constructing the artificial languages to be used in Experiment II. CHAPTER II EXPERIMENT I Six pronounceable nonsense syllables were selected as artificial language elements: faw, ged, hib, jat, mep and zir. The Archer (1960) association value of these syllables is as follows: few, 52%; ged, 36%, bib, 44%; jat, 41%, mep, 36%; and zir, 31%. It seemed possible that certain syllables might be associated with others prior to any exposure to the experimental situation. If associations were to be established through experimental training, any difference in pre-experimental associative strength would make data interpretation difficult. To determine the extent to which the syllables were asso— ciated with each other, a multiple-choice questionnaire was designed and administered to naive subjects. The multiple-choice format in which §S are limited to a set of alternatives has been used in natural language word association tests (Crown, 1947; Kjeldergaard, 1962; Malamund, 1946; Maller, 1936; Terman and Mills, 1936; Wynne, Gerjuoy and Schiffman, 1965). Method Four counterballanced forms (Appendix A) of the twelve item multiple-choice test were given to a group 21 22 of 421 undergraduate students in Introductory Psychology at Michigan State University. Along with the question— naire, each S was also given an IBM Answer Sheet and a scoring pencil. The following instructions were read: You have been given a mimeographed sheet with twelve items. At the top of the sheet there is a Roman numeral I, II, III or IV. Please write this numeral on your IBM Answer Sheet. Please write your name on the IBM Sheet and at the tOp of the mimeographed sheet. The S paused for a moment while the Ss carried out these instructions and then continued: If you look at the item labeled Sample at the top of the mimeographed sheet, it says, "The one word that seems to go best with dax is: (1) seb (2) paf (3) nij (4) tez (5) bip." If you thought that seb went best with dax, you would mark or fill in the one space on the IBM Answer Sheet; if you thought that paf went best with dax, you would fill in the three space; if you thought that tez went best with dax, you would mark the four space; or if you thought that bip went best with dax, you would fill in the five space on the IBM Sheet. Before you start, please say after me the words that you will be seeing in the twelve items. The S then pronounced a syllable and the Ss repeated it. The same procedure was followed for the remaining syllables. Questions asked by Ss were answered by a repetition or paraphrasing of the instructions. The Ss were also urged to complete the questionnaire as quickly as possible. The entire procedure took approximately 30 min. The questionnaires were constructed so that items 1-6 represented one random order of presentation of the six syllables as stimuli, and items 7—12 represented a 23 second. Therefore, each S responded to each syllable twice. The number of Ss who responded with each of the five choices was tabulated for each item and for each of the four forms. Results These data indicated that the response did not vary as a function of questionnaire form (see Appendix B), therefore, the data were pooled. The number and per- centage of the Ss giving the various syllable associa- tions on presentations 1 and 2 are presented in Table 3. Each of the six syllables had frequency distributions of responses which differed significantly (p < .001) from chance distributions on both the first and second pre— sentations. These data are presented in Table 4. In all cases the syllables which Ss most frequently indicated as associates, i.e., those "seeming to go best with each other," were those which shared common middle vowels. The rank order of syllable associations accord- ing to the mean of the percentage of Ss on presentations 1 and 2 making the association is shown in Table 5. Discussion The percentage of Ss indicating that certain syl— lables "seemed to go best" with each other was used as index of associative strength. The syllable association highest in associative strength was "mep" and "ged," while 24 TABLE 3.--Frequency and percentage of associates to six nonsense syllable stimuli on presentations 1 and 2 of the syllable questionnaire. Presentation I Presentation 2 Stimulus Response Frequency gzht Frequency gght faw jat 156 37.2 151 36.3 Zir 110 26.3 111 26.7 ged 73 17.4 70 16.8 mep 51 12.2 49 11.8 bib 29 6.9 35 8.4 ged mep 177 42.2 170 40.7 bib 77 18.4 66 15.8 jat 66 15.8 83 19.9 zir 55 13.1 50 11.9 faw 44 10.5 49 11.7 hib Zir 151 36.0 150 35.9 ged 95 22.7 102 24.4 mep 67 16.0 63 15.1 jat 63 15.0 62 14.8 few 43 10.3 41 9.8 jat faw 128 30.5 121 28.9 zir 98 23.4 101 24.2 ged 90 21.5 110 26.3 mep 54 12.9 48 11.5 hlb 49 11.7 38 9.1 mep ged 200 47.7 178 42.6 bib 80 19.1 60 14.3 zir 49 11.7 63 15.1 jat 47 11.2 56 13.4 faw 43 10.3 61 14.6 zir hlb 146 34.8 128 31.0 jat 103 24.6 101 24.4 faw 73 17.4 87 21.1 ged 66 15.8 59 14.3 mep 31 7.4 38 9.2 25 TABLE 4.-—Va1ues of X2 obtained for the frequency distri— bution of responses to each of six nonsense syllables on presentations 1 and 2. Syllable Presentation 1 Presentation 2 faw 120.18 108.88 ged 136.47 120.26 hib 83.58 ' 88.74 jat 51.11 68.40 mep 211.28 147.69 zir 88.80 59.83 26 TABLE 5.-—Rank order of nonsense syllable associations according to mean per cent on presentations 1 and 2. Rank Stimulus Response Mean Per cent 1 mep ged 45.2 2 ged mep 41.4 3 faw jat 36.8 4 hib zir 36.0 5 zir hib 32.9 6 jat faw 29.7 7 faw zir 26.5 8 zir jat 24.5 9 jat ged 23.9 10 jat zir 23.8 11 hib ged 23.6 12 zir faw 19.2 13 ged jat 17.9 14.5 faw ged 17.1 14.5 ged hib 17.1 16 mep hib 16.7 17 hib mep 15.5 18 zir ged 15.1 19 hib jat 14.9 20 mep zir 13.4 21 ged zir 12.5 22 mep faw 12.4 23 mep jat 12.3 24 jat mep 12.2 25 faw mep 12.0 26 ged faw 11.1 27 jat hib 10.4 28 hib faw 10.0 29 zir mep .3 30 faw hib .6 27 the syllable association lowest in associative strength was "mep" and "zir." Using the indices of pre-experimental associative strength, two artificial languages were de- signed and appear in Table 6. TABLE 6.-—High and low adjective association--artificial languages Stimulus Response Size Nonsense Shape High Adjective LOW Adjective Association Association 1 W1 mep zir zir few 1 X1 mep faw zir ged 3 Y3 jat ged jat mep 3 Z3 hib ged hib mep One of the languages contained adjective elements in which the pre-experimental associative strength was high, i.e., mep and ged, while the other language contained adjective elements in which pre-experimental associative strength was low, i.e., mep and zir. An attempt was made to match the pre-experimental associative strength of an adjective syllable with each of the two noun elements with which it was to appear in the artificial languages. CHAPTER III EXPERIMENT II Once the artificial languages were designed, Experi- ment II was run. Method Subjects Thirty-two undergraduate students enrolled in the Introductory Psychology course at Michigan State Uni- versity were used as S5. There were 20 females and 12 males. The students served as Ss as part of the course requirements. Two Ss were replaced during the running of the experiment because they failed to reach criterion during Session 1. Apparatus and Materials The apparatus used was a combination tachistoscope and memory drum. The Ss looked into the apparatus and viewed materials presented on an automatic card changer. The stimulus shapes were photographic copies of modified historical maps. Each of the four shapes was produced in three different sizes. The "small" size shapes were 28 29 approximately 2.5 sq. in., the "middle" size shapes were approximately 3.75 sq. in., and the "large" size shapes were approximately 6.5 sq. in. The four small size shapes appear in Figure l, the four middle size shapes appear in Figure 2, and the four large size shapes appear in Figure 3. Each of the shapes was mounted in the center of a bluegreen 8 1/2 x 11 in. Smead file-divider card. Four reproductions of each of the three sizes were mounted for each of the four shapes. The nonsense syllables, "mep," "ged," "faw," "hib," "jet," and "zir" were printed in 1/2 in. black capital letters on 3 x l in. white gummed labels. Each one of the six syllables was printed on six different gummed labels. Each printed syllable was then mounted in the center of an 8 1/2 x 11 in. file-divider card. The responses which were to be learned during arti- ficial language training were printed on 6 x 1 in. white gummed labels. A total of four of each of the following responses was printed: "mep zir," "mep faw," "jat ged," "hib ged," "zir faw," "zir ged," "hib mep," and "jet mep." Each response was mounted in the center of an 8 1/2 x 11 in. card. Procedure Two paid undergraduate students assisted in running subjects. One of the assistants ran Ss 2, 6, 7, ll, l5, 16, 23, 24, 31 and 32 and the other ran Ss 3, 8, and 28. The author ran the remaining Ss. 3O .flHSEH .pm was cm on: w mcoz mNHm . :SmEm :II.H we: wa .m 31 .H m m m I .Hszu cams oncomcoz oNH :mappaz I .m opswam . = .4 0 I 0 .HHSEHum odmnm mmcmmcoz oNflm :mwpwq:II.m ohsmfim ’ 33 The Ss were run at the same hour on two consecutive days. The procedure consisted of seven phases. Phase 1 - English Word Association Test 1 The main purposes of English Word Association Tests were to adapt S to the experimental situation, and to obtain the associative responses and latencies for natural language stimuli. Once the S was comfortably seated in the darkened experimental room, the S read the following preliminary instructions: In front of you there is a scope-sight. I will be showing you things and timing how long it takes you to say something. During the experiment be sure to say only the one word you want to say, and try to keep your voice at the same level each time you say a word. For practice, look into the machine and say the word you see. The S then presented the word CAT. After S had repeated the word, these instructions were read: From now on do not repeat the word you see in the machine. When you see a word in the machine, say the first word that you think of as fast as you can. Since I'm interested in the first word you say, and how long it takes you to say it, be sure to say just one word and say it as fast as you can. Remember, do not say the word you see in the machine, and try to keep your voice at the same level each time you say the first word you think of. Suppose I showed you the word CAT, (at this point S presented the word CAT) what is the first word you would think of? During the remainder of the English Word Association Test the following twelve English word stimuli were pre- sented to each S in the same random order: dark, slow, 34 lamp, black, sour, salt, man, long, rough, sickness eagle, boy. Each stimulus word was exposed for two seconds. At the same time that a stimulus presentation began, a Hunter timer was started. The timer was stopped manually by S when S made a response. Although a voice key was used, it proved too cumbersome to use. There- fore, the associative latency measure, i.e., time between onset of stimulus word and S's response, included S's relatively constant reaction time. Between stimulus presentations while S recorded the associative response and its latency, the S was performing a button—pushing task. This task was introduced to prevent rehearsal by S between stimulus presentations. Following the prelimi- nary instructions, Ss were then told: To make the task a little more difficult for you, I would like to have you press either of the buttons on your right or left. Do you see the buttons? Keep pressing the button and silently count the number of presses you make. I will ask you what number you are on, then I will say "Ready?" You will look into the machine and I will show you a word. As soon as you see the word, say the first word that you think of as fast as you can. After you have seen the word and said the first word you thought of, start pressing and counting again. In other words, the procedure will go like this: You press the button and count how many presses you make. I say, "How many presses?" You tell me how many presses you made. I say "Ready?" You look into the machine, see a word, and say the first word that you think of. Then you start pressing and counting again. Do you understand? Just to make sure that you are clear about what you will be doing, could you please tell me what you think your task is? (The S described the procedure.) Please start pressing and counting now. 35 Phase 2 - "Noun only" Artificial Language Training for 16 Trials It should be noted that in the artificial lan- guages generated for the present study, the adjective nonsense syllables are part of the verbal response twice as often as the noun syllables. The possibility that a relationship between frequency of appearance of adjec- tives and frequency of paradigmatic association might operate in the artificial language situation prompted the use of a control for the frequency of appearance of adjec- tives in the present study. The control procedure con- stitutes the "Noun only" Training portion of Experiment II. An item is defined as one stimulus (nonsense shape) and its appropriate verbal response. A trial is defined as the presentation of those four items which constitute the simple artificial language such as those designed for the present study. In a pilot study, it was found that Ss took approximately nine trials to learn simple artificial languages of the High and Low adjective Asso- ciations type. After consideration of (1) the number of times S would be exposed to adjectives as compared to nouns if he reached criterion immediately and (2) the number of trials to criterion required by the group of pilot Ss, sixteen trials were judged to be adequate to balance the frequency of appearance of the adjective and 36 noun syllables. It was also reasoned that in learning a natural language Ss would ordinarily associate objects and their appropriate labels (nouns), and the elements asso- ciated with size (adjectives) would be learned afterward. Therefore, all Ss received sixteen trials of "Noun only" training following the initial English word association test. The stimuli in "Noun only" training, were those nonsense shapes apprOpriate to the particular artificial language S was to learn, but the stimuli were of middle size (Figure 2). The responses were appropriate noun syllables. All Ss were given one "familiarization" trail in which the noun syllables appeared below the middle sized stimulus shapes. The Ss were read the following instruc— tions: Now, let's try something new. I am going to show you some shapes, and each shape will have a nonsense word that goes with it. Your task is to learn the word that goes with each shape. In other words, I will show you a shape and you will say the word that goes with it as fast as you can. When I say "Ready?" you will look into the machine and see a shape and a word printed below the shape. Look at the shape and say the word that goes with it. "Ready?" (S viewed item 1). Look at this shape and say the word that goes with it. "Ready?" (S viewed item 2). Here is the next one. "Ready?" (S viewed item 3). Now look at this shape and say the word that goes with it. "Ready?" (S viewed item 4). At the end of the "familiarization" trial the following instructions were read: From now on you will see only a shape and a short time later the word that goes with it. Try to say the word that goes with the shape before 37 the word appears in the machine. If you make a mistake, or don't know the word, say the word when it appears. Every time you see a shape try to guess the word that goes with it. Even when you say the right word, say it again when it appears in the machine. I will say "Ready?" just before a shape appears. Do you have any questions? Just to be sure that you understand what your task is, please tell me what you think will be happening. Each stimulus was presented for a two second interval, and approximately one second later the response was presented for two seconds. The inter-item interval was also two seconds. At the end of four trials the S had to stop for about two minutes to replace the stimulus and response cards on the card changer of the exposure device. Thus, the inter—trial interval varied from two seconds for four consecutive trials to approximately two minutes between every fourth and fifth trial. The S's response was scored as an error if (1) the appropriate syllable was not given, or if (2) the response was not made within the two second stimulus interval. The number of errors made during the sixteen trials of "Noun only" training was tabulated for all Ss. Phase 3 - Simple Artificial Language Training Following "Noun only” training, sixteen randomly assigned Ss were given training on the High Adjective Association Artificial Language, and sixteen randomly assigned Ss were given training on the Low Adjective Asso— ciation Artificial Language. The languages learned by 38 each S appear in Appendix C. All Ss were run to a cri- terion of three perfect trials. Again a response was scored as an error if (1) the correct two-syllable re- sponse was not given, or if (2) the response was not made within two seconds. The inter- and intra-trial relation— ships remained the same as those of "Noun only" training. Each S received a "familiarization" trial in which the apprOpriate responses consisting of two nonsense syllables appeared below the stimulus shapes, and was instructed as follows: Now let's try something new. I am going to show you some shapes, and each shape will have two non— sense words that go with it. Your task is to learn the two words that go with each shape. In other words, I will show you a shape and you will say the two words that go with it as fast as you can. When I say "Ready?" you will look into the machine and see a shape and two words printed be- low the shape. Look at the shape and say the two words that go with it. "Ready?" (S viewed item 1). Look at this shape and say the two words that go with it. "Ready?" (S viewed item 2). Here is the next one. "Ready?" (S viewed item 3). Now look at this shape and say the two words that go with it. "Ready?" (S Viewed item 4). At the end of the "familiarization" trial the following instructions were read: From now on you will see only a shape and a short time later the two words that go with it. Try to say the two words that go with the shape before the words appear in the machine. If you make a mistake or don't know the words, say the two words when they appear. Every time you see a shape try to guess the two words that go with it. Even when you say the two words correctly, say them again when they appear in the machine. I will say "Ready?" just before a shape appears. Do you have any 39 questions? Just to be sure that you understand what your task is, please tell me what you think will be happening. Once the S had demonstrated that he understood the in— structions, the S began the presentation of the stimuli and responses. Presentation continued until S reached a criterion of three errorless trials. The number of errors and the number of trials to reach criterion were tabu- lated for each S. Phase 4 — Artificial Language Word Association Test 1 After Ss had reached a criterion of three error- less trials, they were given a discrete free association test in which each one of the six nonsense syllables was presented as a stimulus. Each artificial language syl- lable was presented twice, once within each of two random blocks of six. The same syllable order, presented to each S, was as follows: hib, ged, faw, jat, mep, zir; followed by jat, mep, hib, zir, faw, ged. Between stimu— lus presentations the S performed a button-pressing task, as they had during the English Word Association Test, to prevent rehearsal. Each stimulus appeared for a two second interval, and Ss were instructed as follows: If you remember when you first came in I timed how fast you said the first thing you thought of when I showed you a word. We're going to do that again with the words you have been learning, that is, the words that went with the shapes. As you did before, press the button and count silently. I will ask you how many presses you made, then I 40 will say "Ready?" and show you one of the words that went with the shapes. Say the first word you think of - one of the new words you learned - as fast as you can. Do you understand? Just to be sure, could you please tell me what you think will be happening? The associative response and latency were recorded for each of the twelve syllable stimuli. After completing the Artificial Language Word Association Test, the Ss were reminded to appear at the same time the following day, and the first session ended. Phase 5 - Continued Simple Artificial Language Training At the start of the second session of the experi- ment, the Ss were given additional training in the simple artificial language. The procedure was the same as that of Phase 3 with two exceptions: (1) no "familiarization" trial was given; and (2) training continued to a criterion of six errorless trials. The number of errors and the number of trials to criterion were recorded for all Ss. Phase 6 — Artificial Language Word Association Test 2 After Ss had reached the criterion of six error- less trials, they were given a discrete free association test in which each of the six artificial language non— sense syllables was presented as a stimulus. The order of presentation of syllables was as follows: faw, ged, jat, mep, zir, hib; followed by mep, jat, faw, zir, hib, ged. The procedure was identical to that used in the 41 previous Artificial Language Word Association Test of Phase 4. The associative response and latency of each of the twelve syllable stimuli were recorded. Phase 7 - English Word Association Test 2 Following the Artificial Language Word Association Test, all Ss were given the same English Word Association Test as that of Phase 1. The Ss were instructed as follows: Now, we are going to do the very same thing, except that the words I am going to show you are the English words. Press the button and count silently. I will ask you how many presses you have made, then I will say "Ready?" and show you an English word. Say the first word that you think of as quickly as you can. Any questions? The associative response and latency for each of the twelve English words were recorded. The Ss were thanked for their participation, and dismissed. An explanation of the experiment was given to all Ss during one of their class meetings after all data had been collected. In summary, the procedure consisted of the follow- ing phases: 1. English Word Association Test 1. 2. "Noun only" Artificial Language Training for sixteen trials. 3. Simple Artificial Language Learning (Noun- Adjective) to a criterion of three perfect trials. 42 4. Artificial Language Word Association Test 1. (End of Session I) 5. Continued Simple Artificial Language Learning (Noun—Adjective) to a criterion of six per— fect trials. 6. Artificial Language Word Association Test 2. 7. English Word Association Test 2. (End of Session II) CHAPTER IV RESULTS Artificial Language Word Association Test The Artificial Language Word Association Test was administered after each S reached a criterion of three errorless training trials, and again after each S reached a criterion of six errorless trials. Since each artificial language contained six syl- lable elements, when presented with a single syllable stimulus, S might respond with any one of the five remain- ing syllables. When an adjective syllable was presented as a stimulus, S might respond with: (l) the other adjec- tive element (Adjective—Adjective paradigmatic associa- tion); (2) one of the two noun syllables which appeared contiguously with the adjective stimulus during training (Adjective-Noun contiguous syntagmatic association); or (3) one of the two noun syllables which had not appeared contiguously with the adjective stimulus, but had occupied the same contextual position as the adjective (Adjective- Noun positional syntagmatic association). When a noun syllable was presented as a stimulus in the word associa- tion test situation the S might respond with: (l) the other noun syllable which occupied the same contextual 43 44 position and had also appeared contiguously with the same adjective syllable during training (Noun-Noun positional paradigmatic association); (2) one of the other two noun syllables which had not occupied the same contextual posi- tion as the stimulus noun (Noun—Noun non-positional para— digmatic association); (3) the adjective syllable which had appeared contiguously with the stimulus noun during training (Noun—Adjective contiguous syntagmatic associa- tion); or (4) the adjective syllable which had not appeared contiguously with the noun stimulus, but which occupied the same contextual position during training (Noun— Adjective positional syntagmatic association). The possible associations which could have been made by Ss in each of the Artificial Language Groups, and the classification of the associations in terms of paradig- matic or syntagmatic, as well as contiguous, positional, or non—positional appears in Table 7. Within a single word association test, each non- sense syllable was presented twice. Therefore, Test I consisted of Presentations 1 and 2, while Test 2 con- sisted of Presentations 3 and 4. Data analyses were done in terms of Presentations rather than Tests, because of the possible influences of the first presentation on the second presentation within each test. Therefore, although data obtained on second presentations were analyzed and presented, the value of these data is limited. It should 45 see pmH smH zen see as: use as: mm m 2H: and Q2: pan. m» m new new now see 3mm LHN Hx H new paw 3mm new new LHN H: H see she has smH has use maezm «New DH; new new 2H: one LHN omcoawmm msHsaHpm owmzwcmq onEmm 304 awe sue use new see doe paw awe 3mg 2H: one new Q08 9H; pom nos 2H: mos 2H: saw LHN paw LHN 3mg pan LHN 3mm AHN 3mm new has can can new as: AHA saw the son peH see new has mm m new has new DH: Ann new as; new has SH: SH; smH new umH -m» m new 3mm new new 3mm new pew new 9H; LHN new 9H: 3mm nos Hx H one DH: nos 3mg pH: ems 3mm one LHN pan LHN 3mm use new LHN awe H3 H one new, one LHN new awe LHN one new LHN 3mg LHN new one oomnm oNHm m m m m m m m m m m m m m m omcoomcm msHaeHum HmcoHuHmom msostucoo HmcoHuHm0d mzoszucoo HmcoHuHmoaIcoz HmcoHuHmoa wwmawcwa mHmEmm m>HuowHefisomHe< cwfim csozucsoz m>HpuoHeHuomHe< cme mCOHumHoomm< OHumeprczm ozone mcoHpmHoowm< oHumstcwaam ozone .mmmmswcmH HmHoHernm :H moms mCOHumHoommm no COHpmlomemHoII.n mqmH.H mmo.o m H m T Hm.H mmo.o m m 00.0 000.0 0 o sm.H mmH.o : 2 m H o mm.H mmH.o OH m oe.m m:o.o m m HA.H :mH.o m m H m 0H.H moo.o : m 0H.H szo.o m m om.H mmH.o o m a m u“ No.0 300.0 a m mo.H mmo.o m m mm.H me.o o e m in Hm.H :mo.o : m ::.H 300.0 a m mm.H moo.o m H m H u. mw.H mmH.o s m mH.H mmo.o m m mm.H moo.o m m H m m . m m m m m m m m m m :OHpmp pmme mo mo mo wcH>Hm mo mo mo mnH>Hw mo mo mo wcH>Hw Izmmohm .pmq .ooam .02 mm .pmq .oonm .02 mm .pmq .Qonm .02 mm mo .02 go .02 mo .02 dczuz eznz m new mcoHpnoaomQ Um>ammpo w mqm.w: ooN Ho Hm m: m: 9.3 we HH NH :m i H H N >H HHH HH H Emom SCH» Impcmmmhm mmcoammm mzHSEHum a w >H HHH HH H Show COHQ Imucmmmhm mmnoamom wsHSEHpm H w >H HHH HH H Show COHp Imucmmmnm mmcoammm mzHJEHpm 103 m.HH w: MH NH 2H awe N.o: ONH a: w: mm mm awe H.mH mm Hm mH mH HH m QmE m.NH MH 2H NH mH N.N: NNH mm om 2m Nm o.mH No mH :H om mH H LHN m.mH mN mH HN mH m.:H NN NH 4- F N BAH :.mN mm NN mm mm wN w.mH we ON mH OH :H o.mH mm mH om :H HH H m.wN HNH mm Nm NN mN zmH N.HH O: OH OH NH HH 3mm 3mm m.om mNH om Nm ON mm m.OH :: mH :H m.0H m: mH mH w.mH Om OH mH OH MH 9H: m.mm omH Om :: :m N AHN N.HH OH mH mH :.wH NN OH OH NN mH 0.0m HmH mm on mm mm H m.ON OHH Hm OH mm mm wow BH HHH HH H Egon cOHp Impcommnm omcommwm mSHsEHpm N N >H HHH HH H Shom cOHp twpcmmmpm mwGOQmmm msHserm N w >H HHH HH H Epom cOHB Impcmmohm mmcogmmm msHserm APPENDIX C ARTIFICIAL LANGUAGES LEARNED BY EACH S 105 Q08 00w Q08 DH: Q08 DH: O0m LHN O0m LHN Q08 uwm 30H LHN 30m LHN H O Q08 QHQ Q08 umm Q08 pmm 30m LHN 30m LHN Q08 2H5 U0m LHN 00m HHN H O 30% LHN 30Q LHN 60m LHN Q08 DH: Q08 pan 60w LHN Q08 DH: Q08 pmfl m m U0w LHN O0w LHN 30m LHN Q08 umh Q08 OH: 30m LHN Q08 paw Q08 DH: m < Hmm 0mm mmm .mmm mmm 0mm mHm NHm 0NHm mamcm Q08 00m Q08 2H: Q08 2H: O0w LHN O0m LHN Q08 pmh 30w LHN 30m LHN m O Q08 DH: Q08 pmm Q08 00w 30m hHm 30m LHN Q08 9H: O0w LHN 60w LHN m O 30m LHN 30H LHN O0m LHN Q08 9H8 Q08 00H O0w. LHN Q08 DH: Q08 90h H m 60m LHN O0m LHN zmQ LHN 0Q8 paw Q08 DH; 30% LHN Q08 awn Q08 2H: H < mHm :Hm mHm OHM mm Nm mm :m 0NHm mamcm QOOOO mOHBOmmO< 30H 30m Q08 LHN Q08 O0m 2H8 LHN Q08 30m Q08 O0w paw 00m pmm O0m pHn H O LHN Q08 30Q Q08 O0m p0m 30m Q08 LHN Q08 O0w DH: O0m DH: O0m paw H O 00w pHc 60m Own LHN Q08 O0m pHg O0m pww BMQ Q08 LHN Q08 30m Q08 m m 60w pmw O0m 9H: 30m Q08 O0w 00m O0w QHO LHN Q08 30H Q08 LHN Q08 m < mmm mmm Nmm 0mm 3mm mmm Hmm mHm 0NHm maHBm 30m Q08 LHN Q08 O0m DH: LHN Q08 30m Q08 00m pmm O0w paw 60m DH: m O LHN Q08 zmQ Q08 O0m paw 30w Q08 LHN Q08 O0m 9H: O0w DH: O0w pmm m O 60m QHL O0m paw LHN Q08 O0m DH: O0w 00H 30H Q08 LHN Q08 30% Q08 N m 60w paw O0m DH: 30m Q08 O0m awn 60w DH: LHN Q08 30m Q08 LHN Q08 H < mHm MHm HHm mm mm mm mm Hm 0NHm mamBm QDOEO mOHBOmOQ< mOHm APPENDIX D RAW DATA: ARTIFICIAL LANGUAGE WORD ASSOCIATION TESTS 107 .HO.H mm.o :N.H mm.o ON.H OO.H mm.o OO.H OO.H HH.H OO.N mm.o mm.o HH.H OO.H mm.H .p0H oz 02 02 oz 02 02 oz 02 02 oz 02 oz 82 02 a < mmmHO LHN 30H 30H 30% 30M 30m HHN LHN 30H 30m 30m LHN 9H3 3mm O0m 60m :8 N m.m W... om.H mm.o OO.H ON.H mm.H Nm.H Om.H Nm.H Hm.H mo.H mm.o HH.H om.H Nm.m .pmH umme 82 oz 02 oz 02 02 a 02 02 02 oz oz 82 02 a < mmeO H: we 30H 30m 309 LHN O0w LHN 30H HHN 30H 30m 00H 30H O0m O0w mm m:.H Nm.o OO.N :H.H OO.H Om.o Hm.H OO.H ON.N mOHH OO H N0.0 mO.H Nm.o O:.H ON.H .pmq 02 oz 82 02 02 02 oz 02 oz 02 oz 02 02 02 a 02 mmmHQ QHN LHN DH: 30% 30m 30m LHN LHN HHN 30m LHN 30m HHN 30H 60w 3mm mm :m.N oz 02 02 oz 02 02 oz 02 02 < 02 02 oz 02 < 02 mmMHO A.wwmv Q08 HHN NM m 30% N 30m ON 5 mm 30% N LHN HH hHN OH 30% OH 80w mH LHN H 30H O LHN N mm W00 m 30H H Hm m .mSHSEHpm TH dnoas qB, 108 OO.H NN.H Om.O NH.N mo.N NN.N OO.H ON.H ON.H mm.H NH.H HO.H MO.H OO.H OO.H NO.H .90H 02 02 02 oz 82 02 oz 02 oz 02 02 oz mmmHO pmH pH: pH: AHN pan O0m AHN an AHN DH: AHN pH: pH: 9H8 nah pmH Hm Hm.H NH.N No.0 NO.H mz.m om.o Hm.H OO.H mO.H ON.N mo.m mm.o mm.H mm.H No.m mm.m .pmq m same 02 oz 02 oz 02 oz 4 02 oz 02 oz 02 oz 02 02 oz mmmHO 00H 00H pan pan pan pan pHN pan 9H2 00H pan pan pan pmH Sam pm mm Hm.H HO.H OO.H HH.H Hm.m mm.H om.H mm.m mm.m om.H mo.m mm.o mm.H HN.H Hm.m oa.H .004 02 02 oz 82 oz 82 pmnpo 82 02 02 a 02 02 02 < 02 mmmHo 00H 9H8 9H: 30m 00H o0w :GON: SGH 2H8 an nHN DH; pan 2H2 QHN 00H mm C O C C O . O mNHOHNOHNr-INHHHHN \ONNCDCDChOQZ't—lmml-lfim NMINMQDONOQQO=ermN Ln ON .u0H H pmme A.nvmv oz an Hm oz 9H8 om 02 9H; mm a nHu mm 82 cmm mm 82 omw om 82 800 mH a nHN NH a nHN mH oz an HH 82 c0: NH pmnpo =nHH= 0H oz pan m 02 pmH N < MHN O 02 van a mmmHo H8 0 Q08 nmszsHum dnoaa moq 109 .90H 8 02 oz 02 oz 02 02 a 02 02 a oz 82 oz 8 oz mmMHO Q08 DH: 8H8 00H 00H 00H 9H8 Q08 pH: pH: Q08 8H8 pHN 9H8 Q08 8H8 :m N OO.H H0.0 mm.H mm.H mm.H ON.H HN.H OO.H NO.H ON.N NO.H HO.H no.0 OH.H ON.H NO.H .pmq pm0B oz 02 oz 82 oz 02 a 82 82 02 02 a 82 02 a 02 mmmHo pmH 00H 2H8 30m 00H 80H Q08 MHN LHN 9H8 pH: Q08 30m OH: Q08 DH: mm ON.H oz 02 oz 82 oz 82 82 oz 82 oz 82 oz 82 02 a 02 mmmHo 9H8 pmH 8H8 Smh 2H; MHN LHN pmH HHN 9H8 pHN 9H8 30m 9H8 Q08 80H mm H wo.m mN.H NO.H mO.H HN.H ON.N N:.H Nm.: ON.H OO.H OO.H mH.H OH.H 3:.H NO.H NO.H .pmH pmme 82 oz oz 82 oz 82 02 oz 02 oz 02 02 oz 82 02 oz mmmHO A.hO0V O0m pHN Nm pH: mm 9H: Nm 200 mm 2H8 Hm 30% mm 9H8 Hm pan mH DH: OH an mH 9H8 HH DH: m 9H8 m LHN m 9H8 m 8H8 H Hm m ”mSHSEHum dnoas ufitH 110 mm.H no.0 H0.0 NH.H :O.N om.H ON.H :O.N mm.H OO.H H:.H N0.0 NN.H ON.H ON.H H:.H .pmq o< ca o< ea ea ca 08 08 ma ca o< 0< 0< o< 0< 0H wwMHO hHN LHN QHN LHN Q08 mHN LHN pHN Q08 LHN nHN LHN mHN LHN HHN aHu :m N ON.N O0.0 OO.H No.0 mo.N Om.o Om.o NN.H mN.H mm.o ON.H mm.o Nm.N OO.H HO.H NO.H .umq pm0B mmmHO LHN hHN 9H2 hHw AHN LHN Q08 HHN HHN HHN AHN Q08 30H Q08 AHN LHN mm m:.H ON.H OO.H ON.H :O.N ON.H Om.H mo.N mm.H mO.H OO.H No.0 OO.H :m.H :O.H NO.H .pmH o¢ o< ch o< SH 0¢ 0¢ o< 0< 0< SH 0< 0< 04 ch o< mmmHO HHN HHN DH: QHN Q08 QHN LHN HHN HHN hHN Q08 mHN HHN pHN 00H AHN mm H NNOLflO-S'ONLO mOHa-oxOmON N r-lt—‘ll—it—lr-ir-lv—INr—IOI—i LON LDKO Hm.o OO.H OO.H OO.N OO.H .umq pm0B o¢ 0< 82 0< d< 0< 0< ch ch 0¢ 0¢ 80390 04 0< 0< o< mmmHO Assosv O0w LHN Hm pHN om 300 mm pHN mN Q08 mN HHN om nHN mH 00H NH pan mH mHN :H nHN NH :QmE: OH nHN m HHN N LHN O LHN H Hm m "mSHSEHum dnoaa moq 111 0< o¢ o< o¢ 0< 0< Sz 0< 0< 0< 0< o< Q82 o¢ 0< 04 wmmHO Q08 Q08 Q08 Q08 Q08 Q08 30H Q08 Q08 Q08 Q08 Q08 SSH Q08 Q08 Q08 :8 OO.H NO.H ON.H :O.H OO.H mN.H OO.o OO.o mm.H NO.H ON.N OH.H OO.o mH.H NO.H Om.H .pmq m pmme 0< 0< 0< 0< 0< Sz o¢ o< o< 0< 0< 0< ch o< o¢ o< mmmHO Q08 Q08 Q08 Q08 Q08 30m Q08 Q08 Q08 Q08 Q08 Q08 9H8 Q08 Q08 Q08 mm NO.H ON.H Hm.H mN.H mH.H no.N om.o mm.H om.H mo.H m:.H HH.H OO.H Om.H ON.H HN.H .pmq mmeO 60w Q08 Q08 30m 30m pan Q08 Q08 Q08 Q08 Q08 Q08 OH: SSH Q08 Q08 mm :O.H mO.o NN.H O:.H Nm.H O0.0 NN.H OO.H m:.H mN.H ON.H OO.H OO.H HO.H :N.N mm.H .umq H p009 SH SH o< SH Sz ca o< SH o< SH SH ch SH 08 SH mmmHo Ansocv SHN omw mm Q08 ON Q08 NN Q08 ON 309 :N Q08 NN Q08 HN Q08 OH Q08 OH Q08 OH Q08 HH Q08 80H O0w Q08 Q08 Hm mkanunuvn ”03H38Hum dnoaa USTH 112 o o 9 o o NHHNHHHOOI—IH OH~GDOP~OV¢P~ Ner anmNOKOJONCIDUDM ON.H mN.H NO.H OO.H mO.H .pmH 02 oz 02 02 oz 02 02 oz 02 02 02 oz 02 < 02 oz mmmHO O0w 30m 30% 30m 30m 30H 30m O0w 30H U0w 60w saw 30m Q08 30m O0w 2m QNmMMWHNON: moooo—a-Hmzo-zrmm Hr-{r—lr—Ir—lI—lr—lt-lr—lr—lt—i :- N H Hm.H 02 oz 02 02 02 02 < 02 oz 02 oz 02 oz 02 oz 02 mmMHO ©0w 60w 30H 3mm 30m 30m Q08 O0w 30% 60w 30m 30H 3mm 30m O0w 60w mm OO.H HO.H OO.o ON.H ON.H mm.H am.H OH.H OO.H OO.N mN.H :O.H mm.H mm.H NN.N Hm.m .pmH 02 S2 Sz 02 oz 02 02 a 02 oz 02 02 02 a 02 oz mmmHo U0w DH: 00H 30m 30m 60w 30m Q08 30m 30H 30% amp 30m Q08 U0w 60m NS H OO.H NO.o :H.H H:.H m:.H ON.H HO.H OO.H HO.H OO.H OO.H HN.H NO.H 2:.H mm.N OO.H .umq pm0B 02 oz 02 oz 02 02 02 ¢ 02 02 oz 02 02 < 02 oz mmmHO A.mev SHN cmm Hm Smw om 36% mm 30m mN 30S mN 00w om 30% OH Q08 NH 30S mH 3mm 3H 30m NH 30S OH 30S O Q08 N U0w O U0m h Hm m "mSHBEHpm dnoao M01 113 o:.H mO.o zOO NO.o :N.H NO.H ON.H OO.H OO.H O3.H OO.H OO.H OO.H OH.H ON.H OO.H .pmH 0< 0< 0< 0< 0< o< SH 0< 0< o< 0< 0< ch o< o< o< mmeO O0w 60m 60m U0m 60w 60w Q08 60w O0m U0w O0w 00w LHN 60m U0w O0w Hm o< Omw o< O0w o< U0m o< O0w o< O0w o< ©0w ch 30H oH U0w o< U0w hwfipo :me: oH 60w o< O0m ch LHN o< 60w o< 60w o< U0w mmmHo mm Nm.H mm.o Hm.H mm.H mo.H ON.H SO.H HN.H Sm.H Nm.H Hm.H HH.H Hm.H OO.H omH m:.H .pmq oa oa SH 0< 0a 0< 0a 0< o< SH ea 08 S2 0< 0< 0a mmmHo 60w O0m Q08 O0m 60m O0m Q08 O0w O0m Q08 O0m O0w 9H3 O0w O0w O0m mm H OHOOKO HHHOt-‘l ON.N HO.H Nm.H ON.H :O.H om.H OO.o H0.0 OO.H OO.H HN.H .pmH pm0B C\O[‘N\D o< o< o< o¢ 0< 0< 0¢ 04 0< 0< o< o< Sz ch 0< 0< mmmHO Assocv pan 80m mm cmw mm Sow Nm 60m ON O0w :N Smw mm 3mm HN O0w OH O0w OH emm mH Smm HH 60m O 9H3 m HHN m umw m 60w H Hm m "03H58Hpm dnoaa qBIH 114 OO.H ON.H OO.H OH.H HO.H OH.H ON.H OH.H OH.H OO.H OO.N OO.H ON.H mN.H NN.H NO.H .umq 0< 0< o< o< SH SH SH SH SH SH SH SH SH SH mmmHo Q08 Q08 Q08 Q08 LHN Q08 Q08 Q08 LHN Q08 LHN Q08 Q08 Q08 Q08 Q08 Hm N :rN<3 cno:: HrHr4H OO.H mm.H NN.H HO.H :O.N NN.H NN.H H0.0 NO.H OO.H ON.H OO.N .pmH pm0B SH SH SH SH SH SH SH SH SH SH SH SSSSS SH SH SH SH wmmHo Q08 Q08 Q08 Q08 Q08 Q08 Q08 Q08 LHN Q08 Q08 :DmE: Q08 Q08 Q08 Q08 mm NO.H OH. NO.H HH.H ON.H O0.0 NO.H OH.H OO.H OO.H OO.H NO.o NN.N OH.H NH.H ON.H .pmH SH ,SH SH SH SH SH SH SH SH SH SH SSHSS SSHSS SH SH SH mmeO Q08 Q08 Q08 Q08 Q08 Q08 Q08 Q08 SHN Q08 Q08 :DmE: :pON: Q08 Q08 Q08 mm H om.H OH.N OO.o OO.H NO.H HO.H OH.H OO.H Hm.N OH.N NN.H HO.o ON.H OO.H :O.N HO.H .pmq pm0B o< mmmHO Assocv awn Q08 Hm DH: om Q08 ON Q08 ON Q08 mN Q08 ON SHN OH AHN NH SSH mH SH: HH SSS NH Q08 OH Q08 O Q08 N Q08 O Q08 h Hm m "msz8Hpm dnoas moq 115 OO.H HH.H Hm.H ON.H .pmq SH SH SH SH SH SH ch SH SH SH SH SH Sz SH SH SH SSSHO USO USO USO USO USO USO HHS USO USO USO USO USO SSH USO USO USO Hm NH.H O0.0 OO.H HOHH HN H ON.H OOHO mN H HO.H NN.H HN.H OO.H H0.0 NO.H NO.H OO.o .pmq m SSSS SH SH SH SH SH SHO SSS .H OH. SH SH SH SH Sz SH SH SH mmmHo USO USO USO USO USO USO :>HS= USO USO USO USO USO SSH USO USO USO mm \0 OK!) CD N-Z' 0083mm magmam Or-iu—it-i 21' . H SOMOHHHH r-i r-ir-i H OO.o NN.H ON.H .SSH SH SH SH SH SH SH SH SH SH SH SH SH ch SH SH SH SSSHO USO USO USO USO USO USO USO USO USO U0: USO USO HHS USO USO USO mm Hm.H SHHH mo H mSHH NO.H OH.H NH.H OH.H SS H mSHH mS.H SH H MH.H mm.H mm.H SS.S .SSH H SSSB Scpo 9 SH SH SH ch SH SH SH SH ch SH SH ch SH SH SH SSSHO Assocv QHQ DON: MW Scmw m SSS Nm USO Om 30m HN SSS mm USO HH USO OH USO NH HHS H SSS H USO O HHS m SSS m SSS m USO H Hm S .S3H58Hpm H dnoaa qBI 116 SH SH SH SH acz SH SH SH SH SH SH hSnpo SH SH SH SH SSSHS QSE QSE SSE QSE 3mm QSE QSE QSE QSE QSE QSE : 908: QSE LHS QSE QSE Hm mmHNLflwN moomooxom r-lr-iNt-lOt-{r—i SH SH SH SH Sz SH SH SH SH SH SH SH SH SH SH SH mmmHo QSE QSE QSE QSE SSH SSE pHN QSE SSE QSE QSE SSE QSE SSE QSE QSE mm SS.H SS.H NH.H SS.H SS.H HN.H Sm.H HS.H S:.H Sm.H SS.H SS.S So.H Sm.H HH.H mm.m .SSH SH SH SH SH 922 SH SH SH SH SH Q22 mSnpo SH SH SH SH SSSHQ QSE QSE SSE QSE USO QSE QSE QSE QSE QSE USO : DmE: QSE QSE QSE QSE SS H HH.H NN.H mH.H :o.m mm.H ON.H HH.H mm.H HN.H OO.H mo.m mm.o Hm.H m:.H NH.H mo.m .pmq pmSB SH SH SH Scz ch SH SH SH ch SH ch SH SH SH SH SSSHQ Assocv nHz SSE Hm QSE om QSE mm SSS mm SSS mm QSE om QSE mH QSE SH SSS mH HHS HH USO NH SSE 0H QSE m SHS S SSE m QSE h HS S “SSHSEHSO dnoag M01 117 Um.H HO.H mm.o mm.o om.H om.H mm.o mw.H mm.H mm.o NN.H Hm.H No.0 mm.H mm.H mm.H .pwq SH SH SH SH SH SH Sz SH SH SH SH SH acz SH SH SH mmwao QSE QSE SSE SSE SSE QSE AHN QSE QSE QSE QSE SSE SSH SSE SSE QSE Hm mS.H SS.H SS.S HN.H HS.H SS.H mz.H HO.H Sm.H SS.H MH.H OO.H HS.S Sm.H OH.H SS.H .SSH m SSSB SH SH SH SH SH SH SH SH SH SH SH SH ch SH SH SH SSSHQ QSE QSE QSE SSE QSE QSE USO SSE QSE QSE SSE QSE DH: SSE QSE QSE mm Sm.m mS.S SS.H SH.H mm.H Hm.H Hm.H NN.H OH.H SS.S mm.H HH.H Hm.H mH.H mm.H SS.H .SSH SH SH SH SSSSS SSSHS QSE QSE QSE zoSQ: QSE USO USO USO QSE QSE QSE QSE AHS QSE QSE QSE mm OO.N OO.H OH.H OO.H mm.m mm.H mm.H mm.m mm.H HN.H mm.H MH.H mm.H OH.H NO.H mm.m .pmq H SSSS SH SSE mm SH QSE mm SH SSE Sm SH SSS Sm SH SSE Hm SH SSS mm SSSSS =SHH= Hm 9Hz SH: SH SH SSE UH SH QSE MH SH SSE HH SH SSE w Scz SSH m SH QSE m SH SSE m SH SSE H SSSHS Hm S ”SSHSEHSO Assocv 3mm dnoas qSIH 118 OH.H SS.S SSHS SS.H SS H mm.o mm.H mS.S Hm.H SS.H mS.H SS.H SS.S SH.H SS.H Hm.H .SSH SH SH SH SH SH SH SH SH SH SH SH SH SH SH SH SH wwwHD HHS HHS HHS HHS HHS HHS HHS HHS HHS HHS HHS HHS HHS HHS HHS HHS Hm mw.H mm.o Sm.o mo.H OH.H mm.o mm.H NN.H wN.H mm.H mm.o mm.o mm.H HN.H ON.N mm.H .pSH S SSSH SH SH SH SH SH SH SH SH SH SH SH SH SH SH SH SH mmmHO HHS HHS HHS HHS HHS HHS HHS HHS HHS HHS HHS HHS HHS HHS HHS HHS HS om.m mm.H om.H mm.H mo.m mH.H mm.H mm.H HN.H mm.H mm.H mm.o mm.H mH.H mm.H OH.H .HSH SH SH SH SH SH SH SH SH SH SH SH 922 SH SH SH SH SSSHS HHS HHS HHS HHS HHS HHS HHS SSE HHS HHS HHS SH: HHS HHS HHS HHS SS mm.H mm.H :w.o mm.H ©N.H mo.m NH.H Hm.m mmHH mm H HN.H HH.H mw.m om. mm.H mm.H .HSH H SSSH SH SH SH SH SH SH SH ch acz SH SH SH SH SH SH SH SSSHS Assocv BSH HHS Hm HHS o HHS mm HHS mm HHS mm HHS SS HHS aw SHS NH SSH H HHS H HHS SH HHS SH HHS m HHS S HHS m HHS h HS S .mSHSEHum dnoaa M01 APPENDIX E FISHER'S EXACT TESTS OF NUMBER OF SS GIVING AHA ASSOCIATIONS IN THE HIGH AND LOW GROUPS A—A Other A-A Other A—A Other A—A Other 120 High Low 2 5 7 14 11 25 16 16 32 High Low 1 2H *5 15 12 27 l6 16 32 High Low u l 5 12 15 27 l6 16 32 High Low 5 5 10 11 11 32 l6 16 32 .3942 .3326 .3326 APPENDIX F THE x2 TESTS OF FREQUENCY OF A-A ASSOCIATIONS FOR HIGH AND LOW GROUPS ON EACH OF THE FOUR PRESENTATIONS 122 Presentation 1 High Low A-A Rs Yes 2 6 8 No 30 25 55 32 21 63 Presentation 2 High Low A-A Rs Yes 2 A 6 No 30 27 57 32 31 63 Presentation 3 High Low A-A Rs Yes 6 2 8 No 26 30 56 32 32 65 Presentation A High Low A-A Rs Yes 6 5 11 No 26 27 53 32 32 64 2.55 0.22 1.28 APPENDIX G RAW DATA: ERRORS PER TRIAL FOR "NOUN ONLY" TRAINING ining Number of Errors per Trial for Noun Only Tra Trials 124 l6 15 1A 1 3 O O O O O O O r—iOr—{OOr-‘lm OONOHOOOOOr—IONO l3 HMOOOOOOOt—IOI—IOOON OOHHOOOOOOOt—ir—IO 12 l l O l O O O O 0 0000000 OOOOOOHOOHOOI—IO ll 10 9 O 2 O O 0 O l l l 2 l l l l O 3 O O l l O 2 0 0 2 l O l 2 O 8 l 2 O l O 2 O O O 2 O l l O l 3 O O O 2 l O 2 l l l O 2 l O MUMNNNHNNHNNOHOMN HONNmm-S'HHHHNNO NSNNNMNNNNMHONH-fi': HONMZ‘NMNNNNNMI—i t—ILII'MNNNNNNNMMI—INNMJ MNNMMJJNNMNJMN mll—Immmoor—imxooor—i rifirarim 10 12 14 l5 l7 19 2O 23 25 28 30 3 31 APPENDIX H RAW DATA: ERROR PER TRIAL FOR ARTIFICIAL LANGUAGE TRAINING TO CRITERIA 1 AND 2 Trials to Criterion 2 Criterion 2 ARTIFICIAL LANGUAGE TRAINING Trials to l 2 3 A 5 6 7 8 9 10 ll l2 13 14 15 l6 l7 l8 19 2O 21 l 2 3 A 5 6 7 Number of Errors Per Trial 126 Nr—ir—lMONr—IOMLOMMNOOO r-i O H O r-! O H H H (MOI-IO NOr—IOONr—IOMMNNHOOO N m m m H o H o H m m N H H .4 H 0 cm 0 O 2 l N m H H r-{r-i H H NH 00 t—th—i (\l H r-IO m Or-IO (\l H NO Hm NNHHm O (Ur-i om MMNONt—IN MNN r-l(\l t—INNOMOr-‘l HNON om NJMNMHHNHNHMNHHM :TNNHMONNMMNNNOOM JNMHMNNMMSNMNOJJ t—INMLCNGDHMKOQDI—{N-Z'KONO NNNN (\l r—‘In—ll—IHNN m Trials to Criterion 2 Criterion 1 (Artificial Language Training, Cont.) Trials to 1 2 3 A 5 6 7 8 9 10 11 12 13 IA 15 16 17 18 19 2O 21 l 2 3 A 5 6 7 Number of Errors Per Trial 127 \ozxomxomn-Hmmmzmor—im H H OH O OH H t“. OH C) H 1"! HO Nt—l NNN r-ir—IN r-lr—i OO MNIMNNt—iI—INr—INNNOr—lm N O N N N N H H H N r-i H N H N N N H H H H N :—l H r—{u—i O NN N H N v—ir—i t—i NO H H m mo r-i man—{m N N r—lOr—i N MHON N m MNr-i JHZ’t—lt—ir—l mmm NMNH NOMHHONMHHHHMHO Ht—IJMMNHMOJHMSOH MOMNJHN-ZTMMMHMNON NMt—INMHI-‘IJN-Z'm-U'MZ’MN :TKDb-ChON-Z'mNChOMLflwOr—i Hr-lt—lr-lr—ir-{NNNNMM APPENDIX I RAW DATA: ENGLISH WORD ASSOCIATION TESTS 129 :H.H mm.H 9mg Hm.H mm.H ow.H mm.H mm.H mm.m mo.H m».H 5:.H Hm.H mm.H H:.H mm.H NN.N NN.H mm.H mm.H mm.o :m.H :m.o mm.H mm.o om.o NH.H m:.H m:.H HO.H mm.H m>.H om.H mm.H umq spooEm SpooEm spoosm :sooem Chm: cmSOE smsoe zozom spooEm spoosm mmmm mmmm cpoosm spoosm xoom spooEm nwzoe zuOOEm zuooEm Upmx spoosm cpooEm zaEzm zQEzm cmzoe EmSOE Luoosm spoosm boom Spoosm spooEm zmzoe cwsoe pmom bum: ammm Lpoosm xmmm ohm: cuOOEm Hm NN.H om.fi mm.o 3:.H No.0 _mo.H mm.o om.H mm.m NN.H OO.H mm.H mm.o No.0 :O.H sw.fi ON.H 0H.H 00.0 mm.H ma.m om.H HO.H m:.H mo.H HO.H om.H mH.H mq.H oH.H mm.H No.0 umq NN.H H:.H NO.H mz.H mm.o mm.H 2H.m mm.H mm.H mm.o HO.H OO.H mm.H mm.H Hm.> HN.H mm.H mH.H Hm.H mm.m mo.H H0.0 Hm.H 0H.H mo.H H:.H no.0 um.H NO.H Hm.H mm.o 6mg psocm psogm upogm swam OZOH ppogm psogm spocm 30am psocm upocm upocm smog sogswz psogm psocm usocm psogm psosm psozm pnozm pbozm “Loam “nozm psocm spasm upozm steam uponm steam uposm psocm psocm ptogm unocm spasm psogm Hm :m.H mm.H mm.o Hw.o mH.H wo.H mm.H mm.o NN.H mm.H NO.H mm.o ON.H mm.o NO.H mm.H mm.H wo.H mm.o mm.o mm.H HN.H om.H HO.H mo.H wo.H m:.H no.0 mm.H :o.H mm.H wo.H pmq mm.H mm.H mm.H H:.H sm.H mm.H no.0 Hm.H :s.a mm.o mo.H :Q.H OO.H NN.H ms.m NN.H mm.m mm.a mm.a mo.m ON.H mm.o :m.o OH.H mo.H mo.a :O.H m:.H mo.H No.0 mm.o pad mozoomm 2H mmHoszahqonfiJrunukao 131 (English Word Associations and Latencies in Seconds, Cont.) R1 Bird Clock Fly Bird Bird Hawk Hawk Bird Hawk Bird American Bird Bird Hawk Bird Bird Bird Bird Golden Bird Bird Bird Bird Bird Tall Bald Bird Bird American Bird Bird Bird EAGLE R2 Hawk Flying Bald Bird Bald White FHOFJHF‘FHAFJHPJRJHFJCDHCDRJHFJFHHFJFWJFJHFJ \o H HPJR) LA) U1 1.44 FJHFAFMDFJOFAFHHFJOFAFHHFJPrehw4k4m+4kuHFdwteanFJH IUJ \OCIDNO‘xKfl-II'UUMH 132 (English Word Associations and Latencies in Seconds, Cont.) Girl Girl Girl Girl Girl Girl Girl Girl Girl Girl Girl Girl Girl Girl Girl Girl Dog Girl Girl Girl Girl Girl Girl Girl Girl Girl Girl Girl Girl Girl Girl Girl BOY Child Play Lat. OHHHHHHOHHNHHHHHHOHHHHHHOOO HHOH .97 .75 .26 .00 .39 .20 .12 .46 .91 .25 .62 .15 .44 .23 .47 .38 .97 .15 .20 .28 .34 .05 .41 .95 .14 .46 .22 HrecuacnocoanrdkwokJOFAHNHFJOCDFHHC>HFAFHHFJFHDFJO 133 mm.a m©.H 00.0 wo.H mo.a NO.H mo.fi No.0 mo.H :E.H NO.H mm.a :m.H NO.H :m.o :H.N mm.H Hm.H mH.H ma.: sm.fi mm.H OO.H m:.a NH.H mo.H mm.a om.o :m.a MH.H mm.H HO.H umd mw.H mm.H mm.m :m.H m©.H mw.H om.H mo.H mm.: m>.H m:.H mm.m 5:.H HN.H mw.H H:.H mo.m mm.H :o.m mo.H mm.m mm.H m:.m m:.m pm.H NN.H H:.H 2m.H mo.H :m.a OH.H Ema paws; memsm pzmflq mason psmfiq pgmflq dz:.H mm.H mm.H 2:.H m:.H mm.H sm.H m:.H m:.H m:.H Hm.H mm.H mm.H 0H.H :N.H mm.H HH.H m©.H HH.H mm.H Fm.H NO.H mm.H mo.H N©.H mm.o Ema wozoomm 2H mmHozmedd Qz< mZOHE