language and the brain Injury to the brain caused by stroke or trauma can disrupt a person's ability to speak, to understand language, or both. Language disorders known as aphasias, resulting from neurological damage, led researchers to examine the relationship between language and the brain. Clinicians and scientists throughout history have observed that language impairments generally result from damage to particular areas of the left cerebral hemisphere, and have suggested that specific parts of the hemisphere are involved in producing and comprehending language.

Early studies of aphasia

One of the most celebrated proponents of this view was the French surgeon Paul Broca. In 1865, Broca published a case report of a patient known as ‘Tan’, who had suffered a stroke that damaged the inferior frontal gyrus of the left hemisphere (now referred to as Broca's area). This region is adjacent to areas of the brain that control the movement of the facial muscles and muscles of speech. Tan had a severe impairment in the production of speech; he was able to utter only a single word, ‘tan’, which he pronounced with great effort. Broca concluded that this brain region was responsible for the production of speech sounds. The aphasia characterized by hesitant, effortful speech, but relatively normal comprehension of language, is now referred to as Broca's aphasia.

Shortly after Broca published his study of Tan, German psychiatrist Carl Wernicke reported cases of two patients with damage to the left posterior temporal lobe adjacent to the primary auditory area of the left hemisphere (now referred to as Wernicke's area). These patients were able to speak fluently, but the language they produced was nonsensical, and they had great difficulty understanding what was said to them. Thus, unlike Tan, these patients could produce the movements of speech, but were unable to comprehend or formulate spoken language. Wernicke proposed that they had lost the sound patterns associated with spoken language. Wernicke called this disorder ‘sensory aphasia’, and claimed that the damaged region of the temporal lobe was responsible for storing the sound representations of speech. The aphasia characterized by fluent, disordered speech and severely impaired comprehension is currently referred to as Wernicke's aphasia.

The apparent anatomical separation between the production and comprehension of speech led Wernicke and others to suggest that language might be represented in the brain in terms of a set of discrete ‘centres’, each with its own specific function. In 1885, Ludwig Lichtheim proposed that spoken language was subserved by three interconnected centres: a motor centre (M) responsible for the movement patterns of words; an auditory centre (A) for the sound patterns of words; and a conceptual centre (C) for the meanings of words and the formulation of ideas. According to Lichtheim's Connectionist Model of language (which was elaborated by Norman Geschwind in the twentieth century), each of these centres was associated with a specific brain area. Damage to an area, or to a connection between two areas, would produce an aphasia. Lichtheim's model not only accounted for the aphasias reported by Broca and Wernicke, but it predicted the existence of aphasic disorders that had not yet been reported. For example, the model predicted that damage to a connection between the auditory and motor areas of the brain would result in an inability to repeat spoken words, without difficulty in producing or understanding them. This disorder, called conduction aphasia, was later observed in a number of patients and is associated with damage to the arcuate fasciculus, a tract of fibres that connects Broca's and Wernicke's areas.

Although the Connectionist Model provided a compelling and powerful account of aphasic disorders, the approach was criticized by a number of clinicians and researchers in the late nineteenth and early twentieth centuries. French neurologist Pierre Marie studied numerous cases of speech impairment resulting from neurological damage, and observed that many patients with apparent Broca's aphasia had suffered damage to regions other than Broca's area, and that damage to Broca's area did not always result in classical Broca's aphasia. British neurologist John Hughlings Jackson criticized the methods and assumptions of the Connectionist approach, and argued that ‘to locate the lesion which destroys speech and to locate speech are two different things.’ Hughlings Jackson also questioned the model for its separation of language from other cognitive functions, and proposed that language and intellectual ability are inextricably linked. A similar view was adopted by British neurologist Henry Head, who favoured the investigation of both linguistic and non-verbal abilities as a means of examining the nature of language impairments in aphasic patients. Despite these criticisms, the conception of the brain as a set of discrete centres continues to influence models of language to the present day, and concerns about the localization of function and the relationship between language and more general cognitive processes remain active debates in current research.

Clinical and psychological approaches

Clinical studies of aphasic patients over the past century have attempted to establish a more detailed description of the language impairments observed in particular aphasic groups by systematically examining various aspects of language processing in individual patients. Comprehensive batteries of tests have been developed to provide quantitative measures of fluency and comprehension and to assess a range of linguistic abilities. The classical disorders of Broca's and Wernicke's aphasia are now well-defined clinical syndromes. Broca's aphasia is characterized by a disturbance in the articulation of speech with comprehension relatively spared, coupled with a phenomenon known as agrammatism: individuals with Broca's aphasia tend to omit or substitute grammatical function words and markers in their spontaneous speech, and they have difficulty producing and comprehending complex sentences. In contrast, Wernicke's aphasia is characterized by fluent articulation coupled with severely impaired comprehension. The speech of Wernicke's aphasics tends to be nonsensical, and typically includes jargon (utterances that are not actual words in the individual's native language, e.g. ‘habe’) and semantic paraphasias (substitutions of a word of similar meaning for an intended word, e.g. ‘cat’ for ‘dog’). The clinical definitions of the syndromes of Broca's and Wernicke's aphasia have formed the basis for subsequent research into the nature of language processing in these patient groups.

Psychological investigations of aphasic disorders have examined the basic components of language, and have attempted to incorporate advances in linguistic theory into the study of the neural bases of language. A number of researchers have assumed a hierarchical model of language, in which linguistic information is processed in a series of discrete stages. The first stage in the hierarchy is the processing of speech, which involves the discrimination and identification of speech sounds (speech perception) and the planning and implementation of articulatory movements (speech production). Sheila Blumstein has examined the perception of speech sounds in aphasic patients and found that many individuals with aphasia have difficulty discriminating and labelling speech sounds. However, Blumstein and her colleagues have not observed a consistent relationship between deficits of speech perception and higher-level language comprehension impairments, including the comprehension deficit observed in Wernicke's aphasia. This finding calls into question the classical account of Wernicke's aphasia as an impairment in the ability to perceive the sounds of speech. Blumstein has also examined speech production in aphasia, and has observed that Broca's aphasics are impaired in the timing and co-ordination of speech-motor movements, whereas Wernicke's aphasics are impaired in the planning of sequences of speech sounds.

The next stage in the hierarchical model of language processing is the ‘mental dictionary’ or lexicon. The lexicon is thought to contain information about words, including word form (the sequence of sounds that uniquely characterizes a word) and lexical-semantics (the meaning of the word and its relationship to other words in the lexicon). Many aphasic patients have difficulty comprehending word meanings and producing word forms. For example, Wernicke's aphasics and other fluent aphasics tend to produce semantic paraphasias in their spontaneous speech, and have difficulty judging semantic relationships between words and sorting words into meaningful categories.

These observations have led researchers to suggest that Wernicke's aphasics may be impaired in the processing of lexical-semantic information. In addition, certain types of aphasics have particular difficulty producing the names of objects, and these naming deficits can be specific to particular classes of objects, such as animals or tools. Hannah Damasio has found that patients who are impaired in naming animals tend to have suffered damage to brain regions associated with the processing of finely detailed visual information. In contrast, patients who are impaired in naming tools tend to have suffered damage to the intersection of brain regions associated with visual processing and the implementation of motor movements. The results suggest that the perceptual properties of particular objects may have influenced the neural organization of the mental lexicon: animals may be identified primarily on the basis of their visual features, whereas tools may be identified in terms of both their functional properties and their visual characteristics.

The next stage in the hierarchical model is the grammar, which is responsible for the production and comprehension of sentences. Grammatical relationships between words in a sentence can be expressed in two ways: syntax (the order that words occur in a sentence) and morphology (grammatical markings on the words themselves). A number of researchers, including Harold Goodglass, Alfonso Caramazza, Edgar Zurif, and David Caplan, have systematically examined the production and comprehension of grammatical information in Broca's aphasia. These investigations have revealed that Broca's aphasics are impaired in the generation and interpretation of morphological endings (e.g. forming the past tense ‘jumped’ from‘jump’) and complex syntactic structures such as passives (‘The boy was kissed by the girl’) and embedded clauses (‘The boy that the girl kissed is tall’). However, Broca's aphasics have no difficulty using word meaning to determine who did what to whom in sentences such as ‘The apple was bitten by the boy’, indicating that Broca's aphasics are able to use word meaning, but not syntax, to interpret complex sentences. David Caplan has proposed that there is a component of working memory that is responsible for deriving meaning from syntactic structure, and that this component is damaged in Broca's aphasia.

Theoretical perspectives

The results of these studies suggest that damage to particular brain regions may result in the disruption of specific aspects of language processing. Linguist Noam Chomsky proposed that the brain contains a ‘language organ’ that is specialized for the representation of linguistic information. This claim has been extended by theorists such as Steven Pinker, who have proposed that language is organized in terms of a set of domain-specific modules, each of which may be associated with distinct neural mechanisms. This approach is supported by the apparent separation between the lexicon and the grammar suggested by the disorders of Wernicke's and Broca's aphasia. However, recent studies have revealed that the language impairments associated with particular aphasic groups may not be as clear-cut as was originally believed. For example, when a variety of different measures are used, lexical and grammatical impairments are typically observed in both Broca's and Wernicke's aphasias.

A number of theorists have suggested that the apparent domain-specific language deficits in aphasia may in fact reflect impairments in more general language processes. For example, Sheila Blumstein and William Milberg have proposed that the language impairments observed in both Broca's and Wernicke's aphasia result from disturbances in the activation of words in the lexicon, which may also interfere with the construction of sentence-level meanings. Further, Elizabeth Bates has observed that the syntactic structures that are most commonly impaired in aphasia are also disrupted in neurologically intact individuals under general cognitive stress, suggesting that grammatical impairments do not necessarily result from damage to a grammar module. It is hoped that future research into the nature of the language impairments observed in aphasic patients, coupled with advances in brain imaging techniques, will shed new light on the relationship between language and the brain.

Jennifer Aydelott Utman

Bibliography

Caplan, D. (1992). Language: structure, processing, disorders. MIT Press, Cambridge MA.
Sarno, M. T. (1991). Acquired aphasia. Academic Press, Inc., San Diego CA.

See also speech; stroke.