Much of the research noted in this paper was supported by NASA under grant NCC9-58 with
the National Space Biomedical Research Institute.
The Linguistic Review 22 (2005), 223.235 0167.6318/05/022-0223
c Walter de Gruyter
The major “contribution” of generative grammar to cognitive science is negative. The hermetic disjuncture of linguistic research from biological principles and facts has influenced cognitive science. Linguists have followed the pied piper taking a different path from that pointed out by Charles Darwin. As Dobzhansky (1973) noted, “Nothing in biology makes sense except in the light of evolution.” The hermetic nature of much linguistic research is apparent even in phonology which must reflect biological facts concerning speech production. For example, studies dating back to 1928 show that tongue “features” do not specify vowel distinctions. However, the irrefutable findings of these cineradiographic and MRI studies are generally ignored by linguists. Chomsky’s central premise, that syntactic ability derives from an innate “Universal Grammar” common to all human beings constitutes a strong biological claim. But if a UG genetically similar for all “normal” individuals existed, one of the central premises of Darwinian evolutionary biology, genetic variation would be false. Concepts and processes borrowed from linguistics such as “modularity” have impeded our understanding of brain-behavior relations. Some aspects of behavior are regulated in specific localized “modules” in the brain, but current research demonstrates that the neural architecture regulating human language is also implicated in motor control, cognition, and other aspects of behavior. The neural bases of enhanced human language are not separable from cognition and motor ability. The supposed unique aspect of syntax, its “reiterative” productivity, appears to derive from subcortical structures that play a part in neural circuits regulating motor control. Natural selection aimed at enhancing adaptive motor control ultimately yielded a basal ganglia “sequencing engine” that can produce a potentially infinite number of novel actions, thoughts , or “sentences” from a finite number of basic elements. Recent studies suggest that the human FOXP2 gene, which differs from similar regulatory genes in chimpanzees and other mammals, acts on the basal ganglia and other subcortical structures to confer enhanced human reiterative ability in domains as different as syntax and dancing. The probable date of the critical mutations on FOXP2 is coincident with the appearance of anatomically modern human beings about 150,000 to 200,000 years ago. Humans thus can create more complex sentences than chimpanzees, but has anyone ever seen an ape dancing?
The contribution of generative linguistics to cognitive science ultimately may be the realization that the principles and procedures employed by biology are germane rather than those promoted in most current linguistic studies. These include the presence of genetic variation, the opportunistic use of an “organ” adapted for one function for a “new” behavior, and determining what aspects of behavior are “derived” features that may be species-specific. Although Noam Chomsky’s linguistic theories make strong biological claims, they are marked by indifference to biology. Chomsky’s 1957 book, Syntactic Structures presented a linguistic theory that had a biological base; it claimed how the mind works. It presented a refutable, hence “scientific” theory concerning the manner in which humans convey concepts and information. Underlying “kernel” sentences were formed in the mind were postulated that bore a close relation to meaning; “transformational” and phonologic “rules” shaped the semantic form into the flow of speech.
Initial linguistic and psycholinguistic data confirmed the theory’s predictions. For example, psycholinguistic data showed that it took longer for neurologically intact subjects to comprehend highly transformed sentences. The initial, limited, linguistic corpus appeared to be “rule-governed.” However, formal linguistic research started down a slippery slope when psycholinguistic data that refuted aspects of the theory were deemed “production” effects that did not reflect an ideal speaker-hearer’s “competence.” The line between competence and performance is arbitrary. As Bunge (1984) pointed out, it became evident that linguistic theories were tested against a theory of data, data that were arbitrarily thought to reflect “competence,” instead of actual linguistic phenomena. Theories rapidly shifted, guided by vague notions of “simplicity.” Introspection and Occam’s razor, a medieval concept, replaced scientific inquiry. Biological systems conform to the opportunistic, historical path of evolution (Mayr 1982); they are not necessarily logically or economically designed. Nonetheless, linguistic research following the “generative” model adopted strong biological claims such as “modularity” and the centerpiece - the innate “Universal Grammar.”
2. Universal grammar
Ignoring comparative studies of the linguistic abilities of apes which show that ability to talk is the species-specific aspect of human language, linguists focussed on syntax being the unique attribute of human language. The mainstream of linguistic research followed Chomsky, attributing human syntactic ability to an innate organ of the human brain - the “Universal Grammar” (UG). No one would dispute the fact that human beings have an innate biological propensity to acquire language. However, the UG claim is that each and every aspect of the “core grammar” of every human language that was or is spoken is genetically coded in the UG. Hence, the UG must be essentially identical, without significant variation, in every “normal” human being. In short, all of the hypothetical syntactic rules, principles, optimality constraints, and whatever devices are in the current theory, must be present in every human brain (Chomsky 1980, 1986, 1995). In effect, humans are supposed to be superducks whose species-specific vocalizations are genetically coded. Gottlieb (1975) showed that ducklings about to be hatched will produce the full range of adult duck calls after a short exposure to these sounds, without a subsequent “normal” life as a duck. Unfortunately, playing a few tape recordings of Tibetan to your infant won’t result in his or her being able speak Tibetan.
In this regard, recent studies of the effects of Darwinian Natural Selection on Tibetans, argue against a genetically specified UG. Studies of human populations living at sea level reveal a great deal of genetic variation in respiratory efficiency. Some people need to breathe five times as much air through their lungs to transfer the same amount of oxygen to their bloodstream (Bouhuys 1974). Over a span of 40,000 years, Natural Selection for life at extreme altitude with thin air has occurred and “normal” Tibetans as a group are unique; they all can more efficiently transfer more oxygen to their bloodstream. Persons who lacked this genetically transmitted characteristic perished. Moreover, in Tibetan women, placental blood flow to the fetus has been enhanced and newborn infants are about 500 grams heavier at birth than the babies of “lowland” women (Moore, Niermeyer and Zamudio 1998). We can conclude that a period of 40,000 years appears to be sufficient for Natural Selection to enhance innate characteristics that confer biological fitness (the ability to have viable progeny) in human beings.
These findings bear on the question of UG. If an innate UG existed, we might suppose genetic selection that enhanced a child’s “acquisition” of the particular characteristics of the Tibetan language would have been retained while features that were in conflict with Tibetan, but would facilitate the acquisition of other languages would be lost. Thus, if genetic factors played a dominant role in language acquisition we would expect Tibetan children to have difficulty learning a language such as English, which differs markedly from Tibetan. As any linguist at MIT could readily confirm by taking a short bus trip to Central Square in Cambridge, Massachusetts where many children of Tibetan ancestry live, it is evident that they have no difficulty learning to speak English. In fact, no evidence, whatsoever, exists for children whose parents spoke any particular language, having any difficulty learning any other language.
Of course, as is the case for many aspects of current linguistic theory, there is an escape clause for UG - inconvenient facts reflect the “peripheral” grammar. But if Cognitive Science is to remain a science, theories must be subject to refutation.
Modularity represents another dubious transfer from linguistics to cognitive science. According to Fodor (1983), the functional organization of the mind-brain involves a set of modules that instantiate a self-contained process. Modular processing models (e.g., Levelt 1989) propose that in understanding the meaning of a sentence we start with a module that first converts the acoustic signal to a set of discrete phonetic features. A characteristic of modular theory is “shallow information transfer”; the acoustic signal is not transferred from the module that derives phonetic features. Another module accepts the output of this first module and discerns phonemes, a third module recognizes words formed by sequences of phonemes, further modules decode syntax, others feed in semantic information, and so on. However, evidence from studies that date back to the 1960s refutes modular theories for speech perception as
well as “higher” levels of language processing such as word recognition and syntactic processing. Speech produced in normal conversation is generally underspecified. Pollack and Pickett (1963) found that 200 msec segments cut out of the stream of speech were generally unintelligible, even when recordings were made under optimal conditions. The message would suddenly “pop up” as the duration of the excised segment was increased. In some cases the apparent message was incorrect - listeners “heard” words that the speakers had never uttered. The apparent explanation was that listeners formed a running hypothesis on the basis of impoverished acoustic cues; when the hypothesis was consistent with an informed guess based on the listeners’ expectations they “heard” the message, reconstructing nonexistent phonetic detail from their expectations. These findings have been repeatedly replicated (e.g., Tseng 1981; Samuel 1996, 1997, 2001; Shockey 2003). Speakers are sloppy when they believe that a listener can infer meaning from context or prior knowledge (Lieberman 1963). When people talk, their speech output is generally underspecified. Moreover, slight distortions of the acoustic signal that do not have any effect on phonetic categorization, affect lexical retrieval demonstrating that shallow information transfer does not occur in the hypothetical phonetic module (Uttman et al. 2001). In short, little evidence supports the claim that the human brain is organized in a manner similar to a digital computer in which a chain of discrete operations perform particular “encapsulated” calculations. Listeners appear to keep track of the acoustic signal even as they perform ostensibly “higher” operations.
But modular theories live on in Cognitive Science. The most recent claim for Universal Grammar (Hauser, Chomsky and Fitch 2002) entails a species-specific module, a “narrow facility for language” (FLN) that yields “reiteration” - the neural process whereby a finite number of words can be recombined and reordered to generate an infinite number of sentences. The FLN confers recursive syntax devoted to language, and language alone. However, data from studies that follow the paradigms of evolutionary biology - that includes comparative studies of the brains and behavior of a wide range of species show that the neural basis for reiteration is not limited to syntax, nor is it strictly a species-specific human attribute. It has evolutionary antecedents that go back hundreds of millions of years in time.
4. Neural circuits
The key to understanding the nature and evolution of the probable neural bases of human reiterate ability starts with the realization that Broca’s and Wernicke’s areas of the neocortex are not the “seats” of language. The traditional linguistic neophrenological model transferred to cognitive science is simply wrong. Independent studies show that absent subcortical damage, permanent aphasia does not occur. And as Lashley (1951) proposed a half century ago, evidence from independent studies links the neural bases of human syntactic ability to brain mechanisms that first evolved to meet the demands of motor control in primitive species. The ability to change a motor response pattern when circumstances dictate yielded a neural system that can reorder discrete actions to preserve biological fitness. For example, a frog that is busily feeding will survive when it stops and instead jumps into the nearest body of water at the
approach of a predator.
It is evident that complex behaviors such as my pecking at the keys of this computer, walking, talking or comprehending the meaning of a sentence involve activity in many neural structures linked in a circuit. And within a particular neural structure, one generally finds neuroanatomical structures that perform operations that play a critical role in different aspects of behavior. “Local” operations performed by anatomically segregated populations of neurons in particular structures are linked to neurons in other structures forming neural “circuits” that regulate different aspects of behavior. In a class of cortical-striatal-cortical circuits, the striatal, subcortical, basal ganglia constitute a “sequencing engine” that plays a part in circuits that regulate activities such as walking, talking, comprehending the meaning of a sentence, retrieving the memories and the meanings of words, and changing the train of one’s thoughts. These findings are noted in many independent studies including Cummings (1993), Marsden and Obeso (1994), Graybiel (1995, 1997) and Lieberman (2000, 20002, in press).
Broca’s area is involved in language, most likely being the verbal working memory “space” in which syntactic and lexical information are consolidated as a sentence is comprehended. However, in itself, Broca’s area is not the brain’s “syntax organ.” Current functional magnetic resonance imaging studies (fMRI) studies show that bilateral cortical activity during language tasks is the norm in neurologically intact subjects. Broca’s and Wernicke’s areas, their right hemisphere homologue, many other cortical areas, the basal ganglia and other subcortical structures are activated when neurologically intact subjects comprehend the meaning of a sentence or it’s prosodic content (e.g., Kotz et al. 2003). Moreover, damage to other elements of the circuit such as the basal ganglia can result in deficits in a person’s ability to interpret syntax and to speak similar in nature to those noted in traditional studies of aphasia (Lieberman 2000, 2002) And Broca’s area is involved in manual motor control and the recognition of gestures (Rizzolati and Arbib 1998). Circuits connecting frontal and temporal regions of the brain are also involved in memory (Kirchoff et al. 2000). Relevant studies are discussed in Lieberman (2000, 2002, in press). The point to bear in mind is that the operations carried out in local structures of the brain often are not domain specific or localized. Unfortunately, the modular locationist theory imported from linguistics surfaces in fMRI studies that purport to identify specific parts of the brain that are the “seats” of religion, face recognition, lust, and the like.
5. The FOXP2 gene and being human
The so called “language gene”, FOXP2, is a signal case of how a modular view can lead to misinterpreting critical data. The hypothetical Chomskian UG must, of course, encode the different syntactic schemes that occur in the world’s languages. Therefore, UG must contain many detailed syntactic rules. Since diseases such as diabetes, which have a strong genetic component, result in specific deficits, one source of evidence for UG would be a genetic anomaly that prevented afflicted individuals from mastering a specific aspect of English syntax, while retaining other aspects of normal linguistic ability. This was reported to be the case for the afflicted members of a large extended family (KE) who suffer from a genetically transmitted anomaly. Gopnik (1990) and Gopnik and Crago (1991) claimed that these individuals’ linguistic deficits were specific - limited to their being unable to master specific syntactic rules, the regular past tense of English verbs and regular plural nouns. Pinker (1994) repeated and publicized these claims. Other aspects of English syntax, cognitive and motor behavior supposedly were similar to the normal members of family KE.
However, this is not the case. FOXP2 is a regulatory gene that governs thedevelopment of the subcortical neuroanatomical structures of neural circuits that regulate motor control, mood, cognition, and language (Lai et al. 2003). A point mutation on this gene in the family studied by Vargha-Khadem and her colleagues (1998) results in anomalous basal ganglia and other subcortical structures that support neuronal populations that project to Broca’s area, motor and other cortical areas. Intensive study of family KE reveals a “syndrome” - a suite of severe speech and orofacial movement disorders, cognitive deficits, and linguistic deficits that are not limited to specific aspects of the syntax of English (Lai et al. 2001, Vargha-Khadem et al. 1998, Watkins et al. 2002). Major orofacial sequencing errors occur. The afflicted members of family KE are not able to stick their tongues while closing their lips; they have difficulty repeating two words sequences. On standardized intelligence tests, afflicted members of family KE have significantly lower scores than their non-afflicted siblings, which rules out environmental factors that might affect intelligence. Some of the afflicted individuals had higher non-verbal IQ scores than unaffected members of the KE family, which leads some investigators to conclude that FOXP2 does not affect intelligence. However, this is misleading because the mean for the affected members was 86 (a range of 71.11) versus a mean of 104 (a range of 84 to 119) for unaffected family members. The difference in means is significant. Moreover, intelligence derives from the interaction of many neural systems and life’s experiences. The range of non-verbal IQs for the non-affected members of the large extended KE family varies and it is impossible to know what the non-verbal IQs of an affected individual would have been, absent the FOXP2 genetic anomaly.
MRI imaging of affected family members showed that the caudate nucleus was abnormally small bilaterally, while the putamen, globus pallidus, angular gyrus, cingulate cortex and Broca’s area were abnormal unilaterally; functional abnormalities were found in these and other motor related areas of the frontal lobe (Vargha-Khadem et al. 1995). The affected neural structures include ones that support cortical-striatal-circuits implicated in talking, comprehending the meaning of a sentnece, thinking and mood (Cummings 1993; Marsden and Obeso 1994; Graybiel 1994; Lieberman 2000, 2002, in press).
This constellation of behavioral deficits results from a dominant point mutation in the FOXP2 gene (Lai et al. 2001). The FOXP2 gene encodes a forkhead transcription factor, a protein that regulates the expression of other genes during embryogenesis. In the case of family KE, the mutation changes a single amino acid, apparently leading to protein dysfunction. Lai and her colleagues (2003) determined the neural expression of FOXP2 during early brain development in humans and of foxp2, its mouse equivalent, in mice. The areas of expression in both the human andmouse brain are similar and include the structures directly involved in sequencing in the human cortical-striatal-cortical circuits that regulate both motor and cognition . the thalamus, caudate nucleus and putamen as well as the inferior olives and cerebellum. These structures are all intricately interconnected. The cerebellum which receives inputs from the inferior olives is involved in motor coordination. The cortical plate (Layer 6) is also affected by the FOXP2 mutation. As Lai et al. (2003) point out, their “data are consistent with the emerging view that subcortical structures play a significant role in linguistic reasoning.”
The evolutionary significance of the FOXP2 gene is becoming significant. A key finding of the Lai et al. (2003) study is the similarity between the mouse and human foxp2 and FOXP2 expression pattern in the developing brain. The degree of similarity between mouse and human brain development suggests its presence in the common ancestor of mice and humans. However, despite the high degree of similarity there are important distinctions between the mouse, chimpanzee and human versions of FOXP2. The techniques of molecular genetics indicate that the human-specific changes in the FOXP2 protein sequence underwent positive selection sometime between 100,000 to 200,000 years ago, which is consistent with independent estimates (Stringer 1998) for the appearance of anatomically modern Homo sapiens - us. The human version of FOXP2 is absent in chimpanzees, who lack the most “derived” species-specific aspects of human language, the ability to talk. Chimpanzees can not voluntarily recombine a finite number of motor control patterns to form a potentially infinite number of words. Chimpanzees also lack complex syntax, enormous lexical ability, human cognitive ability and also can not freely combine a finite number of motor patterns so as to dance. Has anyone ever seen a chimpanzee dancing?
6. Hermetic isolation
One of the fruits of a productive scientific theory is that it relates phenomena that seemingly are unrelated. Newtonian physics, for example, showed that planetary motion and the flight of a cannon ball obeyed similar principles. Newton’s laws of motion” accounted for these seemingly desperate phenomena. But much contemporary linguistic research instead takes a hermetic approach. A narrow range of phenomena and avenues of research are deemed “linguistic,” other relevant data are simply ignored. It is to be hoped that Cognitive Science does not follow the linguistic model. For example, as noted above key elements of the neural circuits involved in motor control also play a part in regulating syntax. No one really knows how biological brains work, but it is clear that algorithmic serial operations generally employed by linguists do not characterize neural processes. Indeed, theoretical linguists know that they have failed at even being able to describe the grammar of any present human language using these algorithmic processes. As Ray Jackendoff, one of Chomsky’s advocates, notes,
thousands of linguists throughout the world have been trying for decades to figure out the principles behind the grammatical patterns of various languages. . . But any linguist will tell you that we are nowhere near a complete account of the mental grammar for any language. (1994: 26)
However, few linguists are ready to accept the premise that although their traditional “rules” may serve as metaphors, they do not provide insight on the true biological bases of human language. Linguists could profit by studying the neural bases of motor control which appears to involve distributed parallel processing. It is my belief that Cognitive Science, unencumbered by linguistic models will provide a better approximation to how brains and minds actually work, which may suggest a neural basis for linguistic “rules”. And it is apparent that biological brains are the result of the messy, opportunistic evolutionary processes that do not necessarily yield elegant, minimally sufficient systems that conform to “simplicity” metrics. Occam’s razor has little relevance in making decisions concerning competing biological models or linguistic theories grounded in biological fact.
Perhaps nowhere is the hermetic nature of the linguistic enterprise evident than in phonology. The system that is commonly used to categorize vowels derives from Melville Bell, the father of Alexander Graham Bell. In the middle years of the 19th century, Bell devised a system aimed at teaching deaf people to talk. Bell thought that the position and shape of the tongue in a person’s mouth was the primary determinant of vowel quality. Bell, of course, had neither radiographs or MRIs available to confirm these hypothetical tongue positions which most theoretical linguists continue to use. “Features”, generally binary, are used to note the tongue positions that hypothetically specify vowels. The position of the tongue in a word is hypothetically lower in the vowel of the word “bet” compared to its position in “bit”; Phonologists typically formulate algorithms employing these features to describe sound changes. Unfortunately, these hypothetical tongue positions do not specify or even differentiate vowels. Over the course of more than 70 years, the shapes and tongue positions that are necessary to produce the vowels of English and a number of other languages have been established in independent studies that used radiographs, cineradiography, and MRI, a few being Russell (1928), Nearey (1979), Baer et al. (1991), Hiiemaee et al, (2002). These studies show that the shape of the tongue body is almost identical for all vowel sounds and is moved about, almost undeformed. Some speakers produce all of the vowels of English (the most intensively studied language), except for the quantal vowels [i], [u], and [a], with the tongue in the same position. These speakers use lip maneuvers (protrusion and constriction) and adjustments in larynx height to generate the formant frequencies that specify different vowels. Moreover, different speakers employ different tongue positions when they produce the same vowel. The traditional vowel height system does not even describe relative differences in tongue height that supposedly distinguish close vowel pairs such as [I] verses [e]. The vowel [I] supposedly is always produced with a higher tongue position than [e], but some speakers produce [e] with a higher tongue position. Speech researchers (e.g., Nearey 1979, Lindbolm 1988) conclude that gradual shifts along acoustic dimensions account for sound changes over time and space. The few linguists who have actually studied ongoing changes in the manner in which words are pronounced across generations, such as William Labov, concur (Labov, Yaeger and Steiner, 1972). The binary tongue features of “mainstream” linguistic research lead to implausible “categorical” jumps along nonexistent, innate specifications of tongue position.
7. End note
The end note here is a paraphrase of Theodosius Dobzhansky’s (1973) observation that, “nothing in biology makes sense except in the light of evolution.” Cognitive science will make little sense until it is viewed in the same light. Variation is a given. Brains are neither logically or optimally designed. Structures that initially performed one function took on new roles, without abandoning their older activities. Thus we can see that the basal ganglia are involved in regulating, walking, manual gestures, speaking, syntax, thought processes as well as regulating mood and personality. Similar apparent disjoint activities mark Broca’s area which is involved in sentence comprehension, the silent speech of phonetic “rehearsal” in verbal working memory, and recognizing gestures through “mirror” neurons. The welcome contribution from linguistics is that language presents a range of species-specific phenomena that we must account for if we are to understand the nature and the evolution of the human brain. The Pied Piper’s song was, and still is, seductive. Cognitive Science can take account of the principles and findings of biology and strive to find out how biological brains really work. These principles and findings hopefully will lead to productive linguistic theories.
Baer, T., J. C. Gore, L. C. Gracco, and P. W. Nye (1991). Analysis of vocal tract shape and dimensions using magnetic resonance imaging: Vowels. Journal of the Acoustical Society of America. 90: 799.828.
Bouhuys, A. (1974). Breathing. New York: Grune and Stratton.
Bunge, M. (1984). Philosophical problems in linguistics. Erkenntnis 21: 107.173.
Chomsky, Noam (1957). Syntactic Structures.
The Hague: Mouton.. (1980). Rules and Representations. Behavioral and Brain Sciences 3: 1.61. (1986). Knowledge of Language: Its Nature, Origin and Use. New York: Prager. (1995). The Minimalist Program. Cambridge MA:MIT Press.
Dobzhansky, T. (1973). Nothing in biology makes sense except in the light of evolution. American Biology Teacher 35: 125.129.
Cummings, J. L. (1993). Frontal-Subcortical circuits and human behavior. Archives of Neurology 50: 873.880.
Fodor, Jerry (1983). Modularity of Mind. Cambridge, Mass: MIT Press.
Gopnik M. (1990). Dysphasia in an extended family. Nature 344: 715.
Gopnik M., and Crago, M. (1991). Familial segregation of a developmental language disorder. Cognition 39: 1.50.
Gottlieb, G. (1975). Development of species identification in ducklings: I. Nature of perceptual deficits caused by embryonic auditory deprivation. Journal of Comparative and Physiological Psychology 89: 387.389.
Graybiel, A. M. (1995). Building action repertoires:memory and learning functions of the basal ganglia. Current Opinion in Neurobiology 5: 733.741.. (1997). The basal ganglia and cognitive pattern generators. Schizophrenia Bulletin 23: 459.469.
Hauser, M. D., Noam Chomsky and W. T. Fitch. (2002). The faculty of language: What is it, who has it, and how did it evolve? Science 298: 1569.1579.
Hiiemae, K. M., J. B. Palmer, S. W. Medicis, J. Hegener, B. S. Jackson, and D. E. Lieberman (2002). Hyoid and tongue movements in speaking and eating. Archives of Oral Biology 47:11.27.
Jackendoff, Ray (1994). Patterns in the Mind: Language and Human Nature. New York: Basic Books.
Kirchoff, B. A, A. D. Wagner, A. Naril and C. E. Stern. (2000). Prefrontal-temporal circuitry for episodic encoding and subsequent memory. The Journal of Neuroscience 20: 6173.6180.
Kotz, S. A., M. Meyer, K. Alter, M. Besson, D. Y. von Cramon and A. Frederici. 2003 On the lateralization of emotional prosody: An fMRI investigation. Brain and Language 96: 366. 376.
Labov W., M. Yaeger and R. Steiner, (1972). A Quantitative Study of Sound Change in Progress. Philadelphia: U. S. Regional Survey.
Lai, S. J., S. E. Fisher, J. A. Hurst, F. Vargha-Khadem, and A. P. Monaco. (2001). A forkheaddomain dene is mutated in a severe speech and language disorder. Nature 413: 519.523.
Lai, C. S, D. Gerrelli, A. P. Monaco, S. E. Fisher, and A. J. Copp (2003). FOXP2 expression during brain development coincides with adult sites of pathology in a severe speech and language disorder. Brain 126: 2455.2462.
Lashley, K. S. (1951). The problem of serial order in behavior. In Cerebral Mechanisms in Behavior, L. A. Jefress (ed.), 112.146. New York: Wiley.
Levelt, W. J. (1989). Speaking: From Intention to Articulation. Cambridge Mass.: MIT Press.
Lieberman, Philip (1963). Some effects of semantic and grammatical context on the production and perception of speech. Language and Speech 6: 172.180.. (2000) Human Language and Our Reptilian Brain: The Subcortical Bases of Speech, Syntax, and Thought. Cambridge Mass: Harvard University Press.. (2002). On the nature and evolution of the neural bases of human language. Yearbook of
Physical Anthropology 45: 36.62. (in press). Towards an Evolutionary Biology of Language. Cambridge Mass: Harvard University Press
Marsden, C. D. and J. A. Obeso (1994). The functions of the basal ganglia and the paradox of sterotaxic surgery in Parkinson’s disease. Brain 117: 877.897.
Mayr, E. (1982). The Growth of Biological Thought. Cambridge, Mass.: Harvard University Press
Moore, L. G., S. Niermeyer and S. Zamudio (1998). Human adaptation to high altitude: Regional and life-cycle perspectives. Yearbook of Physical Anthropology 41: 25.61.
Nearey, T. (1979). Phonetic Features for Vowels. Bloomington: Indiana University Linguistics Club.
Pinker, Steven (1984). Language Learnability and Language Development. Cambridge Mass: Harvard University Press.. (1994). The Language Instinct; How the mind creates language. New York: William Morrow.
Pollack, I., and J. M. Pickett (1963). The intelligibility of excerpts from conversation. Language and Speech 6: 165.171.
Rizzolatti, G. and M. A. Arbib (1998). Language within our grasp. Trends in Neuroscience 21: 188.194.
Russell, G. O. 1928 The Vowel. Columbus: Ohio State University Press.
Samuel A. G. (1996). Phoneme restoration. Language and Cognitive Processes 11: 647.653.. (1997). Lexical activation produces potent phonemic percepts. Cognitive Psychology 32: 97. 127.. (2001). Knowing a word affects the fundamental perception of the sounds within it. Psychological Science 12: 348.351.
Shockey, L. (2003). Sound Patterns of Spoken English. Oxford: Blackwell
Stringer, C. B. (1998). Chronological and biogeographic perspectives on later human evolution. In Neanderthals and Modern Humans in Western Asia, T. Akanyama, K. Akoi and O. Bar-Yosef (eds.), 29.35. New York: Plenum.
Tseng, C. Y. (1981). An acoustic study of tones inMandarin. Unpublished PhD dissertation, Brown University.
Uttman, J. A., S. E. Blumstein and K. Sullivan. (2001). Mapping from sound to memory: Reduced lexical activation in Brocas’s aphasia. Brain and Language 79: 444.462.
Vargha-Khadem, F., K. E. Watkins, C. J. Price, J. Ashbruner, K. J. Alcock, A. Connelly, R. S. J. Frackowiak, K. J. Friston, M. E. Pembrey, M. Mishkin, D. G. Gadian, and R. E. Passingham (1998). Neural basis of an inherited speech and language disorder. Proceedings of the National Academy of Sciences 95: 2695.2700.
Watkins, K. E., F. Vargha-Khadem, J Ashburner, R. E. Passingham, A. Connelly, K. J. Friston, R. S. J. Frackiwiak, M. Miskin and D. G. Gadian. (2002). MRI analysis of an inherited speech and language disorder: structural brain abnormalities. Brain 125: 465.478.
184.108.40.206 - Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR 2.0.50727; .NET CLR 3.0.4506.2152; .NET CLR 3.5.30729; InfoPath.1)