Neurolinguistics, the study of language representations and processes in the human brain, is at the crossroads of fields of science such as linguistics, psychology, neuropsychology, and neurology. The present survey starts with a historical outline and continues with a review of recent studies on mental lexicon and syntax in aphasia. Those studies support the psychological reality of certain linguistic distinctions. We finish by mentioning some future trends in neurolinguistics.

Expressions of distance, like far (away) from the village and 10km from the village, display various syntactic structures. On the basis of Swedish and French data, it is shown that, when such structures appear as locative predicates, certain properties of their heads become problematic, given the requirements of the GB theory of theta-roles. In order to solve these problems, two essential proposals are made: (a) An external θ-role may be used up in the process of establishing a position by means of a path. (b) The subject argument of NP-(copula)-XP may receive an external θ-role of non-lexical origin. Addressing questions of a contrastive nature, the present paper suggests that directive prepositions may assign an external θ-role in Germanic but not in Romance languages.

Section 1 is a general introduction to the theories discussed in the paper, primarily Sperber and Wilson's relevance theory, with its distinction between “implicature” and “explicature”, and Grice's theory of conversational principles and implicature. Section 2 discusses specific types of implicatures that must be approached in a different manner that the implicatures most commonly discussed in the literature. It is argued that some of these implicatures are hard to explain in the context of relevance theory. To this group of implicatures belong those arising from communicating information by repeating information already known to the listener, and from communicating information and humor by ironic understatement. Section 3 examines the division between explicatures and implicatures. Specifically, the definition of this division is dependent on Sperber and Wilson's assumptions about the division between a person's mental lexicon and his encyclopedic knowledge of the world. However, it is shown that the latter division is both philosophically and practically troublesome. Section 4 asks whether it is even possible to define the division between lexical and general knowledge. The conclusion of the present paper is that it is not.

The present study is an analysis of phonologically deviant words in the speech of a group of Norwegin patients suffering from conduction aphasia. The analysis shows that these deviations are not randomly distributed. Their distribution is such that it supports phonological thories which posit hierarchiacal structures both below and above the level of the segment. The deviations can be accounted for within a phonological theory which assumes that a word's lexical phonological representation contains a phonemic level where each phoneme consists of a set of articulatory features which are hierarchically organized with respect to each other, and which also assumes that the phonological representation contains information about the syllable structure of the word.

This case study reports on a patient RS who showed obligatory and automatic access to the constituents of all English compounds. She would thus interpret the compound blueprint as ‘a print that is blue’. She also showed interpretations that suggested a blending of the whole-word and decomposed meanings of compounds. The compound seahorse was interpreted as ‘a small horse that swims’. It is argued that her behaviour presents evidence in favour of a parallel access model of multimorphemic word recognition in which constituents and whole-word forms are simultaneously accessed. Within such a framework, RS's performance can be interpreted as resulting from the inability to suppress the results of morphological decomposition.

Previous studies on the nature of paragrammatism in Finnish have established a continuum of syntactic complexity with normal speech at one end and Broca's aphasics at the other, while Wernicke's aphasics fall between these two groups. In the present analysis, two paragrammatic aphasics are described whose utterances are longer than normal and of a more complex syntactic structure than the utterances in the comparison data. The paragrammatic subjects also make plenty of morphological and syntactic errors. The comparison of the present results with earlier studies on paragrammatism in Finnish shows that paragrammatism is subject to considerable variation

The characterization of agrammatic speech as telegram style is discussed in relation to an exemplifying comparison of linguistic structures and determining activity factors in a Swedish database of agrammatic speech and simulated telegrams. The strong claim that telegram style is a specific register that is also used by agrammatic aphasics does not find support in the data. Features of agrammatic speech that can be in line with a more vague and metaphorical use of the term telegram style are pointed out.

As a result of present-day technological standards, the technique of computer simulation is constantly gaining influence in cognitive science. Neurolinguistics is a special branch of this field in which cognitive capacities connected with language are related to the structure and functions of the brain. It is argued that computer simulation is a useful technique for evaluating neurolinguistic models. This is demonstrated with respect to neural network models of the process of language production.