Grammatical Computation Research Articles

Gradient symbolic representations in Harmonic Grammar

Gradient symbolic representations in Harmonic Grammar

Domain-general and domain-specific computations in single word processing

Language comprehension relies on a multitude of domain-general and domain-specific cognitive operations. This study asks whether the domain-specific grammatical computations are obligatorily invoked whenever we process linguistic inputs. Using fMRI and three complementary measures of neural activity, we tested how domain-general and domain-specific demands of single word comprehension engage cortical language networks, and whether the left frontotemporal network (commonly taken to support domain-specific grammatical computations) automatically processes grammatical information present in inflectionally complex words. In a natural listening task, participants were presented with words that manipulated domain-general and domain-specific processing demands in a 2 × 2 manner. The results showed that only domain-general demands of mapping words onto their representations consistently engaged the language processing system during single word comprehension, triggering increased activity and connectivity in bilateral frontotemporal regions, as well as bilateral encoding across multivoxel activity patterns. In contrast, inflectional complexity failed to activate left frontotemporal regions in this task, implying that domain-specific grammatical processing in the left hemisphere is not automatically triggered when the processing context does not specifically require such analysis. This suggests that cortical computations invoked by language processing critically depend on the current communicative goals and demands, underlining the importance of domain-general processes in language comprehension, and arguing against the strong domain-specific view of the LH network function.

Synchrony and diachrony in Menominee derivational morphology

This paper identifies two difficulties with treatments of derivation in Algonquian languages. In traditional approaches to grammar, in which the morpheme is seen as a unitary entity, morphemes are understood as minimal units of meaning and/or function. Definitions share an appeal to the morpheme’s indivisibility. In the Algonquianist literature, in contrast, some morphemes (‘components’) can themselves contain other morphemes (which we call ‘formatives’) and they can also be synchronically derived from other components or stems. Drawing data from Menominee, we propose that these difficulties disappear if the formatives are seen as historical rather than synchronic units, while the components are the synchronic morphemes. Formatives bear the hallmarks of historical products of morphologization (phonetic/phonological reduction, semantic bleaching, and increase in grammatical function), and we conclude that they are not part of synchronic grammatical computation. This resolves problems present in traditional and modern theoretical approaches to Algonquian derivation, and has broader ramifications for linguistic theory: in both the structuralist and generativist traditions, synchronic grammar has often been seen as expansive, responsible for generating surface patterns that may instead be products of history. This has been the case in phonology and syntax. The present paper provides a study of the phenomenon in derivational morphology, and suggests that a more modest role for synchronic rules is called for.

Grammatical analysis as a distributed neurobiological function.

Language processing engages large‐scale functional networks in both hemispheres. Although it is widely accepted that left perisylvian regions have a key role in supporting complex grammatical computations, patient data suggest that some aspects of grammatical processing could be supported bilaterally. We investigated the distribution and the nature of grammatical computations across language processing networks by comparing two types of combinatorial grammatical sequences—inflectionally complex words and minimal phrases—and contrasting them with grammatically simple words. Novel multivariate analyses revealed that they engage a coalition of separable subsystems: inflected forms triggered left‐lateralized activation, dissociable into dorsal processes supporting morphophonological parsing and ventral, lexically driven morphosyntactic processes. In contrast, simple phrases activated a consistently bilateral pattern of temporal regions, overlapping with inflectional activations in L middle temporal gyrus. These data confirm the role of the left‐lateralized frontotemporal network in supporting complex grammatical computations. Critically, they also point to the capacity of bilateral temporal regions to support simple, linear grammatical computations. This is consistent with a dual neurobiological framework where phylogenetically older bihemispheric systems form part of the network that supports language function in the modern human, and where significant capacities for language comprehension remain intact even following severe left hemisphere damage. Hum Brain Mapp 36:1190–1201, 2015. © 2014 The Authors Human Brain Mapping Published by Wiley Periodicals, Inc.

Specific Language Impairment Across Languages.

Children with specific language impairment (SLI) have a significant and longstanding deficit in spoken language ability that adversely affects their social and academic well-being. Studies of children with SLI in a wide variety of languages reveal diverse symptoms, most of which seem to reflect weaknesses in grammatical computation and phonological short-term memory. The symptoms of the disorder are sensitive to the type of language being acquired, with extraordinary weaknesses seen in those areas of language that are relatively challenging for younger typically developing children. Although these children's deficits warrant clinical and educational attention, their weaknesses might reflect the extreme end of a language aptitude continuum rather than a distinct, separable condition.

Neurocognitive mechanisms for processing inflectional and derivational complexity in English

In the current paper we discuss the mechanisms that underlie the processing of inflectional and derivational complexity in English. We address this issue from a neurocognitive perspective and present evidence from a new fMRI study that the two types of morphological complexity engage the language processing network in different ways. The processing of inflectional complexity selectively activates a left-lateralised frontotemporal system, specialised for combinatorial grammatical computations, while derivational complexity primarily engages a distributed bilateral system, argued to support whole-word, stem based lexical access. We discuss the implications of our findings for theories of the processing and representation of morphologically complex words.

Neurocognitive Contexts for Morphological Complexity: Dissociating Inflection and Derivation

AbstractWe review experimental findings about the processing of morphologically complex words, and provide a neuro‐cognitive account of the mechanisms that underlie this processing. Within the framework of the brain network that supports language comprehension, we argue that complex words created by regular and rule‐based combinations, such as regular inflections in English (e.g. played, jumps), engage a left‐lateralized frontotemporal subsystem, specialized for grammatical computations. In contrast, processing of lexicalized and unpredictable complexity, present in derived words (e.g. bravely, warmth), engages a bilateral subsystem, which underpins whole‐word, stem‐based lexical access. We suggest a potential neurobiological framework for this proposed organization.

The retrieval and inflection of verbs in the spontaneous speech of fluent aphasic speakers

Fluent aphasia of the anomic and Wernicke's type is characterized by word retrieval difficulties. However, in fluent aphasic speech, grammatical deviations have been observed as well. There is debate as to whether these grammatical problems are caused by the word retrieval deficit, by an additional grammatical deficit, or by an integration deficit. Verbs are an interesting word class in this respect, because they are among the words that are hardest to retrieve for many fluent aphasic speakers and some forms require a considerable amount of grammatical computation. For production of a finite lexical verb, the lexical form and inflection for tense and agreement need to be integrated. In the present study, the use of lexical verbs in finite and non-finite form in the spontaneous speech of 8 Dutch fluent aphasic speakers was analyzed. The results show that retrieval of lexical verbs decreases when more grammatical information needs to be computed. It will be argued that this is best accounted for in terms of an integration deficit: the more computation is needed to express a relation between a verb and intrasentential (agreement) and extrasentential (tense) information, the more difficult it is to retrieve a normal range of lexical verbs.

Abstract Grammatical Processing of Nouns and Verbs in Broca's Area: Evidence from FMRI

The role of Broca's area in grammatical computation is unclear, because syntactic processing is often confounded with working memory, articulation, or semantic selection. Morphological processing potentially circumvents these problems. Using event-related functional magnetic resonance imaging (fMRI), we had 18 subjects silently inflect words or read them verbatim. Subtracting the activity pattern for reading from that for inflection, which indexes processes involved in inflection (holding constant lexical processing and articulatory planning) highlighted left Brodmann area (BA) 44/45 (Broca's area), BA 47, anterior insula, and medial supplementary motor area. Subtracting activity during zero inflection ( the hawk; they walk) from that during overt inflection ( the hawks; they walked), which highlights manipulation of phonological content, implicated subsets of the regions engaged by inflection as a whole. Subtracting activity during verbatim reading from activity during zero inflection (which highlights the manipulation of inflectional features) implicated distinct regions of BA 44, 47, and a premotor region (thereby tying these regions to grammatical features), but failed to implicate the insula or BA 45 (thereby tying these to articulation). These patterns were largely similar in nouns and verbs and in regular and irregular forms, suggesting these regions implement inflectional features cutting across word classes. Greater activity was observed for irregular than regular verbs in the anterior cingulate and supplementary motor area (SMA), possibly reflecting the blocking of regular or competing irregular candidates. The results confirm a role for Broca's area in abstract grammatical processing, and are interpreted in terms of a network of regions in left prefrontal cortex (PFC) that are recruited for processing abstract morphosyntactic features and overt morphophonological content.

Restricting grammatical complexity

AbstractTheories of natural language syntax often characterize grammatical knowledge as a form of abstract computation. This paper argues that such a characterization is correct, and that fundamental properties of grammar can and should be understood in terms of restrictions on the complexity of possible grammatical computation, when defined in terms of generative capacity. More specifically, the paper demonstrates that the computational restrictiveness imposed by Tree Adjoining Grammar provides important insights into the nature of human grammatical knowledge.

Restricting grammatical complexity

Theories of natural language syntax often characterize grammatical knowledge as a form of abstract computation. This paper argues that such a characterization is correct, and that fundamental properties of grammar can and should be understood in terms of restrictions on the complexity of possible grammatical computation, when defined in terms of generative capacity. More specifically, the paper demonstrates that the computational restrictiveness imposed by Tree Adjoining Grammar provides important insights into the nature of human grammatical knowledge.

The neural basis of lexicon and grammar in first and second language: the declarative/procedural model

Theoretical and empirical aspects of the neural bases of the mental lexicon and the mental grammar in first and second language (L1 and L2) are discussed. It is argued that in L1, the learning, representation, and processing of lexicon and grammar depend on two well-studied brain memory systems. According to the declarative/procedural model, lexical memory depends upon declarative memory, which is rooted in temporal lobe structures, and has been implicated in the learning and use of fact and event knowledge. Aspects of grammar are subserved by procedural memory, which is rooted in left frontal/basal-ganglia structures, and has been implicated in the acquisition and expression of motor and cognitive skills and habits. This view is supported by psycholinguistic and neurolinguistic evidence. In contrast, linguistic forms whose grammatical computation depends upon procedural memory in L1 are posited to be largely dependent upon declarative/lexical memory in L2. They may be either memorized or constructed by explicit rules learned in declarative memory. Thus in L2, such linguistic forms should be less dependent on procedural memory, and more dependent on declarative memory, than in L1. Moreover, this shift to declarative memory is expected to increase with increasing age of exposure to L2, and with less experience (practice) with the language, which is predicted to improve the learning of grammatical rules by procedural memory. A retrospective examination of lesion, neuroimaging, and electrophysiological studies investigating the neural bases of L2 is presented. It is argued that the data from these studies support the predictions of the declarative/procedural model.

Past tense morphology in specifically language impaired and normally developing children

This study evaluates the input-processing deficit/single system and the grammar-specific deficit/dual system models to account for past tense formation in impaired and normal language development. We investigated regular and irregular past tense formation of 60 real and novel regular and irregular verbs in "Grammatical (G)-SLI" children (aged 9:3 to 12:10) and morphological- or vocabulary-matched younger control children. The G-SLI children and language ability (LA) controls showed quantitatively and qualitatively different patterns of performance. The LA controls, but not the G-SLI children, showed a significant advantage of regular over irregular past tense marking for real and novel verbs. Past tense frequency affected the GSLI children, but not the controls' production of regular verbs, even with stem access controlled for. The G-SLI children's production of regular forms was significantly lower than that of the control groups. Frequency and phonological properties had a similar and significant effect on the G-SLI and LA controls' irregular formation. The G-SLI children's irregular past tense production did not differ from that of the morphological controls, but was lower than that of the vocabulary controls. We argue that the dual mechanism/grammar-specific deficit provides a parsimonious explanation for normal and impaired performance, and suggest that grammatical computations underlying regular past tense formation in normal grammar are impaired (not missing) in G-SLI grammar.

The Sum and Substance of It: The Appreciation of Mass and Count Quantifiers in Parkinson′s Disease

We performed three experiments in order to determine whether patients with Parkinson′s disease (PD) have a sentence comprehension impairment. In the first experiment, we found that nondemented PD patients were impaired at matching a short sentence containing a mass or count quantifier (e.g., "Point to the jar with many") with one of four pictures. Patients erred by pointing to the incorrect mass or count type of substance, but pointed correctly to the large or small amount of a substance in the picture. In the second experiment, we asked the same PD patients to judge short sentences containing mass or count nouns modified by mass or count adjectives (e.g., *"The glass contains many milk"). PD patients were significantly impaired at detecting errors in the agreement between a noun and a quantifying adjective and errors in the appropriate use of the plural associated with a noun. A sentence completion task was performed in order to determine whether patients impaired on both sentence-picture matching and sentence judgment tasks are also compromised in their expression of quantifiers. Some patients impaired on both tasks also had difficulty expressing quantifiers, and it is these patients who may have difficulty performing grammatical computations. Other patients were selectively impaired in their sentence-picture matching or their sentence judgments, suggesting a performance deficit concerned with a particular task. We conclude that sentence comprehension impairments in PD are multifactorial in nature.

Syntax and Machine Translation

Unlike much theoretical linguistics, both the comparative perspective and the automatic aspect of Machine Translation reveal that there cannot be a clear-cut borderline between grammar and lexicon. The interdependence of grammatical computation and lexical storage has become extremely obvious during the research and development work in the METAL-project. After outlining the history of this project, section 1 of this paper sketches the system's projected application environment. The systern's main lexical tools are discussed in section 2, and section 3 deals with its main grammatical techniques. The final section raises a number of problems and issues for further research. Some of these are well-known but still await a satisfactory solution given the present (1987) state of the project. Others are raised much less often and derive from the technical nature of the texts that METAL is intended to translate.