The biological bases of language
Edited by Ovid J. L. Tzeng
Half a century ago, psychologists hardly discussed the issue of language study; instead, they talked about verbal learning and verbal behavior from the associative perspective. The cognitive revolution in the 60’s had given birth to an exciting new interdisciplinary perspective, in which psychologists and linguists were brought together to ask the same questions regarding the nature of human mind. They had developed complementary and potentially synergistic methods of inquiry into the biological bases of language. Indeed, convincing evidence had also quickly accumulated to show that such an interdisciplinary endeavor resulted in fruitful conceptualization of the relationship between cerebral organization and various language functions.
In December of 1993, a group of researchers, consisting of cognitive psychologists, neuroscientists, linguists, and speech scientists, participated in a symposium, which was held in the Grand Hotel of Taipei, to discuss the biological bases of language from the perspective of Chinese language. The Chinese language has what may be the simplest and most austere grammatical system in the world. While most of the world’s languages offer a wealth of different markers on nouns, pronouns, adjectives and /or verbs, Chinese has essentially no verb conjugations and no noun declensions of any kind. Furthermore, although the Chinese grammar does provide for a set of standard word orders, it is also the case that word order can be varies in a number of ways if the speaker wants to emphasize one element more than another. These properties of Chinese grammar raise some fascinating questions concerning the interaction of language-specific features and language uses with respect to various cognitive functions.
In his well-articulated paper, Alvin Liberman set the path for the symposium participants to follow: Is there a linguistic mode (i.e., a specifically linguistic way of doing things) at the level of action and perception? He contrasts the horizontal theory of speech against the vertical theory, and he asks seven questions concerning the organized components of the larger specification for language. He carefully considers how each theory would answer these seven questions and concludes that only the vertical view is appropriate for linguistic processing from a biological point of view.
If one takes the position that speech is special in its linguistic operation, then how script is mapped onto it becomes critical because the view would require that uncovering the special speech code is critical for beginning readers in mastering reading skills. Indeed, in contemporary research on alphabetic reading acquisition, the notion of “phonological awareness” has been a cornerstone in the theoretical construction. In this respect, one should examine the acquisition process of children learning to read the Chinese writing system, which is logographic in nature and thus notorious for its opaque script/speech relationship. Bertelson, Chen, Tseng and Ko carried out a comparative study on this particular problem and they concluded that their results created strong difficulties for the current tendency to minimize differences between the process of word identification involved in Chinese vs. alphabetic reading. Using data from bilingual (Chinese/Dutch) studies, Beatrice de Gelder proposes several alternatives for the relation between phonological awareness and speech processing.
But what is the unit of Chinese word recognition? Characters are orthographic units, which are separated from other writing units by spaces at both sides. Words, on the other hand, are linguistic units, which are listed as an entry in the lexicon with its syntactic, semantic, and phonological information. There have been many contradictory reports in Chinese word recognition and the reason may lie in the inconsistency between Characters and words as unit of perception and reading. Using special experimental paradigms, Hung, Tzeng, and Ho were able to show words were more salient than characters in Chinese word recognition. Such a finding supports the hypothesis that word recognitions mediated by morpheme construction in Chinese. It is important to note that reading a logographic script, such as the one represented by the visually distinctive characters, taps the underlying speech segments and uncovering the effects of the script/speech mapping should give clues to the discussion of the biological foundation of language.
To find out how early childhood experience in acquiring a visually based language (i.e. sign language used among the deaf) modified learners’ perceptual categorization ability in the relevant domains, three experiments were carried out by Klima, Tzeng, Fok, Bellugi, Corina and Bettger with special stimulus displays. That is, stimulus characters were presented one at a time as a continuous dynamic point-light display of movement. It was found that native deaf signers showed superior performance in the reconstruction of the original target. Such an enhanced ability suggests a greater neuronal plasticity in the human brain.
Contemporary discussion of the biological foundation of human language has to deal with the touchy issue of innateness under the theoretical conceptualization of Universal Grammar and modularity. As cogently pointed out in Tai’s paper, “The issue at stake is whether there is a language-specific faculty in our brain/mind that is modularized and independent of other cognitive abilities of human beings. Or, can language faculty be derived from human’s perceptual and general cognitive categories, memory capacity, processing strategies, and conversational structures between hearer and speaker? In other words, the issue centers around whether or to what extent Chomsky’s ‘innateness hypothesis’ is necessary for explaining patterns of human language.”
Tai’s own position on the issue of innateness is quite clear. His careful examinations of the spatial and temporal expressions in the Chinese language let him to conclude that its grammar is, to a great extent, not arbitrary and not autonomous from human’s conceptualization of the physical world. He is able to show that the three important cognitive bases (namely, space and time, categorization, and iconicity) play an important role in Chinese grammar. Kess and Miyamoto reach a similar conclusion based up results obtained from several different popular paradigms, namely, sentence processing in left-branching vs. right-branching language (i.e. English vs. Japanese. Respectively), the reaccessibility of empty categories, and syntactic comprehension impairments exhibited by Japanese aphasics. Indeed, the results suggest that the informative strategies in natural language processing are language-specific and such language-specific differences demand a processing model that would allow for variability rather than absolute uniformity.
The theme of language-specific property is picked up again in Lien and Wang’s paper, in which the unique problem of shape classifiers in Mandarin and Taiwanese is critically examined from a psycholinguistic perspective. The point is clear: Specific language features are essential keys to the discussion of Universal Grammar and learnability.
Ogura’s paper deals with a much profound aspect of the language change. Instead of the much-studied phonological change, she makes a courageous attempt to uncover the universal property of semantic change. It was certainly not an easy job because out understanding of word meaning is much obscure and little orderly knowledge, such as that observed in the case of phonological change, is available. Orgura examined how word frequency and environmental factors act together to determine the leaders and laggers in the metaphoric transfer of sensory terms in English, Japanese and Chinese, and in the development of speech act verbs in English and Japanese and modal auxiliaries in English.
These are all the papers included in the monograph. Together, they highlighted the importance of looking at biological bases of language from the Chinese language perspective. Authors of all these papers make great contributions for our scientific understanding of language processing. They deserve our applause. But for me personally, I thank their patience.
My aim is to ventilate a question about biology of langiage that has long preoccupied me; Is there a linguistic mode–that is, a specifically linguistic way of doing things –at the level of action and perception? In the time allotted to me, I mean to parse that question and invite your attention to its various aspects. To keep question and anther within the bounds of matters I am supposed to know something about, I will limit my discussion to speech in the narrow sense — restriction takes the limitation of my own knowledge properly into account, while nevertheless permitting me, I hope, to speak to the broader purposes of this symposium. It is surely relevant to such purposes that speech in the narrow sense underlies the phonological system that constitutes full half of the dual structure that characterizes all languages, the other half being syntax; and that some of the question I raise about the one half find counter parts in the other.
1. IN THE DEVELOPMENT OF PHONOLOGICAL COMMUNICATION, WHAT EVELVED?
2. HOW IS THE REQUIREMENT FOR PARITY MET?
3. HOW IS SPEECH RELATED TO OTHER NATURAL MODES OF COMMUNICATION?
4. WHAT ARE THE (SPECIAL) REQUIREMENTS OF PHONOLOGICAL COMMUNICATION, AND HOW ARE THEY MET?
5. COULD THE ASSIGNMENT OF THE STIMULUS INFORMATION TO PHONETIC CATEGORIES PLAUSIBLY BE AUDITORY?
6. JUST HOW ‘SPECIAL’ IS SPEECH PERCEPTION?
7. HOW DO SPEAKING AND LISTENING DIFFER FROM WRITING AND READING?
Paul Bertelson; Hsuan-Chih Chen; Chin-Hsing Tseng; Hwa-Wei Ko; Beatrice de Gelder
The notion that the capacity to segment speech into those phonological units that the orthography represents, often called “phonological awareness”, is a critical component of early reading skill has been a cornerstone of contemporary research on alphabetic reading acquisition. We discuss its implications regarding the process of reading Chinese text. After describing the notion, we summarize its main empirical basis, giving special attention to the data from readers of non-alphabetic scripts. We argue that the latter evidence creates strong difficulties for the current tendency to minimize differences between the processes of word identification involved in Chinese vs. alphabetic reading. Finally, a new research project concerning explicit speech representation in Chinese readers is outlined.
While reading acquisition in alphabetic writing systems has been studied extensively, studies of non-alphabetic reading and reading acquisition are much more recent. This paper is about the contribution of the study of reading in a morpho-syllabic writing system like the Chinese to understanding reading acquisition and its disorders. Part I summarizes studies of phonological awareness in morpho-syllabic vs. alphabetic reading and presents new data on onset segmentation ability in bi-lingual Chinese/Dutch speakers. Part II examines data from speech process studies. Part III reviews various alternatives for the relation between phonological awareness and speech processing and suggests aspect of an alternative framework.
Daisy L.Hung 洪兰; Ovid J.L.Tzeng 曾志朗; Chia-yun Ho
There are many contradictory reports in Chinese word recognition research. One of the reasons may lie in the inconsistency between Chinese characters and words as unit of perception during reading. Characters are orthographic units, which are separated from other writing units by spaces at both sides. Words, on the other hand, are linguistic units, which are listed as an entry in the lexicon with its syntactic, semantic, and phonological information. It can not be interrupted by exterior elements, and must be moved as a unit. In most languages orthographic units and linguistic units are isomorphic; such is the case in classical Chinese. However, in modern Chinese, because of phonological changes and expansion of vocabulary, these two units are no longer congruent. A word may be composed of one, two, three, or even more characters. For example, “sky”, “smallpox”, and “ceiling”. Although the vast majority of morphemes consist of one character, these are exceptions. For example, “grape” and “glass”. The purpose of this research is to clarify what is the unit of Chinese word recognition, character or word? Because words are composed of morphemes, we like to know if the type of morpheme construction and number of morphemes will influence Chinese word recognition, and if they do, which is more influential? Most of the previous researches on this topic employed lexical decision experimental paradigm. Unfortunately, due to the confusing criterion of Chinese lexicon, those results are not consistent. The present study adopted Reicher (1969) and Wheeler (1970) word superiority effect paradigm in Chinese to investigate the saliency of characters and words in Chinese word recognition. Three experiments with different morpheme types and morpheme number conditions were carried out. The results showed that words were more salient than characters in Chinese word recognition, and within words of the same number of characters, the morpheme number influences word recognition more than the morpheme type. Words consisting of one morpheme were recognized more accurately than those consisting of two morphemes. In other words, mono-morphemic words were accessed faster than bi-morphemic words. Such a finding supports the hypothesis that word recognition is mediated by morpheme construction.
Edward S.Klima; Ovid J.L.Tzeng 曾志朗; Y.Y.A.Fok; Ursula Bellugi; David Corina; Jeffrey G Bettger
Three experiments were carried out to explore modification inperceptual categorization resulting from early childhood experiencing acquiring a sign language as a primary native language. Subjects included two groups of Chinese children (deaf children whose language was Chinese Sign Language and hearing ones whose language was Chinese) and two groups of American adults (deaf signers of American Sign Language and hearing speakers of English). The stimuli were Chinese pseudo characters, each presented either as a sequence of static fragments or as dynamic point-light displays of the trace of the character form as it is written in the air. Under each condition, the subjects were to reconstruct and draw the pseudo character from the manipulated stimulus. In Experiment 1, Reconstruction from Static Fragments, the deaf signing children showed an enhanced ability to reconstruct the target form, when compared to the hearing children, but only at zero inter-stimulus intervals. In Experiment 2, Reconstruction from Dynamic Point-Light Displays, both groups of Chinese children were asked to reconstruct the target from a continuous dynamic display of movement; the deaf signing children showed superior performance across the board. The third experiment showed that even without any knowledge of Chinese characters, deaf signing Americans also exhibit an enhanced ability to analyze these complex displays of movement compared to hearing English speakers. Results of these experiments highlight some of the basic properties that may be critical for perceptual transfer from one domain to another.
Human language is innate and biologically determined, insofar as chimpanzees and other primates cannot be taught to use human language (Wallman 1992). The issue at stake in contemporary linguistic and psychological theories is whether there is a language-specific faculty in our brain/mind that is modularized and independent of other cognitive abilities of human being. Or, can language faculty be derived from human’s perceptual and general cognitive categories, memory capacity, processing strategies, and conversational structures between hearer and speaker? In other words, the issue centers around whether or to what extent Chomsky’s ‘innateness hypotheses’ is necessary for explaining patterns of human language.
For natives, Chomsky’s innateness hypothesis is necessary in order to explain language universal and learnability in child language acquisition. In Chomsky’s theory of Universal Grammar (UG), both principles of universals (formal as well as substantive), which dictate forms of possible human language, and parameter, which permit variation intra-language or cross-linguistically, are genetically endowed by the language-specific faculty in the human brain/mind. Furthermore, the innateness of UG serves as a metatheory for child language acquisition, in explaining the remarkable rapidity with which the child acquires language despite impoverished input from the adult language. The input data for the child is impoverished because it contains only grammatical sentences of the adult language (the positive data), and little information about ungrammatical sentences (the negative data).
Alternative modes of explanations for language universals have been offered. These include semantic explanations, discourse-pragmatic explanations, processing explanations, and perceptual and cognitive explanations (cf. Hawkins 1988). The negative evidence problem in child language acquisition has been reexamined by Arbib and Hill (1988) and Bowerman (1988), and the results suggest that the complexity in child language acquisition cannot be solved merely by reference to innate UG. Hence, with respect to language universals and learnability, UG does not have as great explanatory value as nativists claim.
The innateness hypothesis also hinges upon the observation that grammatical patterns are not random, and that grammatical rules are structure-dependent. From the point of view of information science, the non-randomness of linguistic structure is a fundamentally important property of human language so that there are possible and impossible patterns of linguistic structures. Chomsky attributes the non-randomness and structure-dependence of human language to the innate language-specific faculty, and excludes the possibility of deriving linguistic patterns from human’s conceptual structures of the physical world. The emphasis of the autonomy of grammar is to underscore the plausibility of the innateness hypothesis. The autonomy thesis maintains that linguistic structures are self-contained and cannot be shaped by human’s experience and conceptualization of the physical world. This autonomy thesis, as pointed out in Tai (1993), has roots in the arbitrariness principle initiated by Saussure and passed on to American structuralists, from Bloomfield to Harris, and on to Chomsky.
In this paper, it will be observed that Chinese grammar is, to a great extent, not arbitrary, and not autonomous from human’s conceptualization of the physical world. The paper focuses on three important role in Chinese grammar, namely: (1) space and time, (2) categorization, and (3) iconicity. They will be highlighted in the following sections.
Joseph Kess; Tadao Myyamoto
A number of far-reaching claims about the nature of Universal Grammar characterize the theoretical literature in linguistics over the past forty years, and they often carry the implicit expectation that language processing mechanisms must also be fundamentally the same. However, the psycholinguistics literature has often been equivocal about the universality of certain key tenets imported from linguistic theory to explanations of sentence processing. This paper examines psycholinguistic results from several revealing paradigms in the literature on English and Japanese, namely, sentence processing in left-branching vs. right-branching languages, the re-accessibility of empty categories, and syntactic comprehension impairments exhibited by Japanese aphasics. Collectively taken, these suggest that the informative strategies in natural language processing are language-specific. While there are a large number of undeniable similarities in the global strategies employed by the processing mechanism, the language-specific differences should make us reconsider our expectations that the processing mechanism manifests itself in one unique way. Theoretical claims about language structure are not always matched by the realities of processing, and our construction of a processing model should allow for variability rather than absolute uniformity. A cognitive science of the mind is more likely to find informative explanations in a limited range of generalized learning and processing principles, applied to language-specific processing strategies, than on a simple reductionist philosophy of science.
Chinfa Lien 连金发; Penying Wang 王本瑛
ABSTRACT This paper present a quantitative study of two sets of sortal shape classifiers carrying respectively the semantic feature of [spherical] and [long] in Mandarin and Taiwanese from a psycholinguistic perspective. The first set of shape classifiers consist of ke棵 and li粒 in Mandarin and liap粒 in Taiwanese, while the second set comprises zhi枝, tiau條 and gen根 in Mandarin and ki枝 and tiau條 in Taiwanese. A psycholinguistic experiment conducted on two groups of 20 subjects reveals in quantitative terms patterns of intersection of shape classifiers in each set. We also compare these two sets of classifiers in Mandarin and Taiwanese in terms of semantic features and try to bring the choice of classifiers to bear on the linguistic background of subjects.
ABSTRACT A great deal of scholarship over the last two centuries has gone into phonological change. With respect to meaning, our understanding is much more obscure. There is not nearly the kind of orderly knowledge on semantic change as there is on phonological change.
The important line of cross-linguistic research that has opened up new perspectives is on how humans categorize the color continuum. Berlin & Kay (1991), Kay (1975), Kay & McDaniel (1978) and Kay, Berlin &Merrifield (1991) have examined the implication of the results for language variation and change. Investigations in this area have revealed some basic tendencies in the color vocabulary of a wide diversity of languages, drawn material from cultures that range from the very simple to the highly technological. Furthermore, the categorization data are directly related to neurophysiologic properties of the visual system. Research along this line, where attention is given to niversal tendencies as well as to their biological bases, offers the best clue toward understanding how semantic change actually takes place.
Williams (1976) pursues this line of research in the transfer of a lexeme from one sensory modality to another, one of the most common types of metaphoric transfer in all languages, drawing very similar evidence from English, Indo-European cognates and Japanese. Traugott (1986, 1987, 1989, 1995), Traugott & Dasher (1987), and Traugott & Ko_nig (1991) explore the general tendency toward greater subjectivity in presuppositional terms, modal auxiliaries and adverbs in Englsih, and speech act verbs in English and Japanese.
However, as Traugott (1987) states, semantic change very rarely applies to items of the same lexical field at the same time, and thus it is impossible to predict either when that will occur, or which lexical item in a given field will change and which will not. Our perspective on this area has been broadened by considering the question within a perspective of lexical diffusion, a process which is implemented in a manner that is lexically gradual, diffusing across the lexicon (Wang 1969, 1976, 1979, 1983a, b, 1987).
In all cases of lexical diffusion, we find leaders and laggers among the words, which raises the issue of what factors determine these changes. One such factor is word frequency. How the interaction between word frequency and environments determines the schedules has been demonstrated in phonological change, e.g., diatone formation in English (Philips 1983), shortening of EModE u# (Ogura 1987, Chapter 5), the early stages of the acquisition of the vowels and consonants in English and German by a bilingual child (Ogura 1990, Chapter 6.) and vowel merger in Shanghai (Shen 1990), in morphological change, e.g., the development of –s in the third person singular present indicative in English (Orgura & Wang 1996), and in syntactic change, e.g., the development of negation in English (Tottie 1991), grammaticalization processes in To’aba’ita (Lichtenberk 1991) and the development of periphrastic do in English (Ogura 1993).
In the present study, we will proceed with Williams’ and Traugott’s investigations, and examine how the interplay of word frequency and environments determines the leaders and laggers in the metaphoric transfer of sensory terms in English, Japanese and Chinese, and in the development of speech act verbs in English and Japanese and modal auxiliaries in English. Furthermore, we will explore the biological bases that underlie these changes.