Lecture at Nice University (January 2019) by Kleanthes Grohmann
University of Cyprus & Cyprus Acquisition Team
Photo by The Roaming Platypus on Unsplash
Kleanthes K. Grohmann is Professor of Biolinguistics in the Department of English Studies at the University of Cyprus (UCY) and the Director of CAT, the Cyprus Acquisition Team (CAT Lab). He received his Ph.D. from the University of Maryland (2000) and has published widely in the areas of syntactic theory, comparative syntax, language acquisition, impaired language, and multilingualism. Among the books he has written and (co-)edited are Understanding Minimalism (with N. Hornstein and J. Nunes, 2005, CUP), InterPhases (2009, OUP), and The Cambridge Handbook of Biolinguistics (with Cedric Boeckx, 2013, CUP). He is founding co-editor of the John Benjamins book series Language Faculty and Beyond and editor of the open-access journal Biolinguistics.
Biolinguistic considerations of the language faculty across phenotypes
Universal Grammar (UG) denotes the species-specific faculty of language, presumed to be invariant across individuals. Over the years, it has shrunk from a full-blown set of principles and parameters to a much smaller set of properties, possibly as small as just containing the linguistic structure-building operation Merge, which in turn has been argued to derive the uniquely human language property of recursion (Hauser et al. 2002)—or rather, the Labeling Algorithm (Chomsky 2012, 2015; Berwick & Chomsky 2016). UG qua human faculty of language is further assumed to constitute the “optimal solution to minimal design specifications” (Chomsky 2001: 1), a perfect system for language. Unfortunately, the human system or physiology does not always run perfectly smooth in an optimal fashion. There are malfunctions, misformations, and other aberrations throughout. The language system is no exception.
This talk aims at considering theoretical and methodological issues that shed light on the human faculty of language with respect to language development and pathology. The main proposal derives from joint work with Evelina Leivada and Maria Kambanaros (Leivada et al. 2017) on the Locus Preservation Hypothesis, which in addition has implications for typical language acquisition, second language learning, and language variation in general. The first part of this talk will build on joint work with Ianthi Tsimpli and Maria Kambanaros(Tsimpli et al. 2017) and present language pathology from the perspective of the underlying system: What can non-intact language tell us about UG? Particular emphasis will be put on evidence from Greek, and how the investigation of impaired (cognitive-)linguistic abilities from one language can inform the study at large—and how it can (not) shed light on the study of a(n impaired) language faculty. The second part picks up the long-standing observation that grammatical markers are not uniformly impaired across speakers of different languages, even when speakers share a diagnosis and the marker in question is grammaticalized in a similar way in these languages. This work aims to demarcate, from a cross-linguistic perspective, the linguistic phenotype of three genetically heterogeneous developmental disorders, such as Specific Language Impairment, Autism Spectrum Disorder, and Down Syndrome.
Boeckx, C. (2006). Linguistic minimalism: Origins, concepts, methods, and aims. Oxford University Press.
Leivada, E. (2015). The nature and limits of variation across languages and pathologies. Doctoral dissertation, University of Barcelona.
Leivada, E., Papadopoulou, E., Kambanaros, M., & Grohmann, K. K. (2017). The influence of bilectalism and non-standardization on the perception of native grammatical variants. Frontiers in psychology, 8, 205.
Tsimpli, I. M., Kambanaros, M., & Grohmann, K. K. (2017). Language pathology. The Oxford Handbook of Universal Grammar, 486-508. PDF
Chomsky’s 1986 Five questions
1. What is knowledge of language? (Humboldt’s problem, Chomsky 1965)
2. How is that knowledge acquired (Plato’s problem, Chomsky 1986)
3. How is that knowledge put to use? (Descartes, bread and butter of linguistics)
4. How is that knowledge implemented in the brain (Broca, imaging, Boeckx)
5. How did that knowledge emerge in the species? (Darwin’s problem, Jewett 1914)
Linguistic theories deals with these questions with a certain form of idealisation (ideal speaker-hearer, idealised monolingual). We have some answers but there are still outstanding issues. This talk addresses a different aspect of these questions, that is how they relate to pathological cases:
1. knowledge of language: Universal Grammar and language faculty
2. acquisition: impact of impairment: delay vs deviance
3. language use: relevance and effect of impairment
4. brain: differential impairment/breakdown
5. species: genetic factors
There is a need for interdisciplinary collaboration.
Chomsky (2005) three factors
- genetic factors (UG)
- epigenetic (experience)
- third-factor considerations (not language-specific but domain-general)
Developmental language impairments (eg SLI) and acquired language disorders (eg aphasia) have been studied for a very long time. For good reasons:
- to inform the diagnosis and therapy of speech disorders
- localisation of language in the brain (imaging, ERP)
- identify which areas of language may be impaired (morphology, phonology, grammar, pragmatics, semantics)
- discover more about the language faculty (FL).
Earlier Universal Grammar (UG) approaches include Government and binding (GB) theory, positing a modular structure of the cognitive system including a language module, which itself has a modular structure. The language module has a hierarchical structure (X-bar theory, theta theory): lexicon, d-structure, s-structure, phonetic form (PF) and logical form (LF). It was thought that in language pathologies, some modules were impaired and that investigating impairment would inform GB theory. More recent research takes a minimalist approach focusing on either a broad or narrow conception of FL.
Language pathology research
Tsimpli, Kambanaros & Grohmann (2017)
UG denotes the species-specific faculty of language, presumed to be invariant across individuals. Over the years it has shrunk from a principles and parameters model to a single operation: “merge.” There may be a neural footprint of “merge” that allows for the recursive properties of language.
Broad language faculty (FLB): all mechanisms involved in language
Narrow language faculty (FLN): subset of mechanisms unique to humans and to language. Modular theory has been abandoned in favour of a minimalist programme.
UG is assumed to constitute the “optimal solution to minimal design specifications” (Chomsky 2001: 1). Yet the physiology does not always function perfectly.
Grodzinsky (1986): aphasics have difficulty interpreting passives (inversion of theta roles). Tree-pruning or truncation hypothesis: difficulties on left periphery could support a modular interpretation of impairment (projection problem). Studies of impairment shed light on invariant human faculty of language.
4 Pertinent issues
- Can pathology affect the core of language abilities (i.e. the use of universal operations and primitives) or does pathology restrict language-specific properties (giving rise to optionality or variability)? (In other words, is there any evidence that language operations such as external or internal Merge are unavailable or otherwise impaired in any population with language pathology?)
- Does language pathology affect linguistic competence (in language-specific options) or does the variability in language use depend on accessing this knowledge due to affected mechanisms mediating language use (such as working memory resources)?
- Does language pathology affect language use differently depending on whether we are dealing with an acquired or a developmental language disorder? (Associated with this question: Could we distinguish between language use vs. language knowledge issues as a function of acquired vs. developmental language disorders?)
- Can we disentangle the contribution of language-external factors in human cognition such as executive control or in the environment, such as input frequency of a particular token, type, or structure from language-impaired performance in developmental or acquired language disorders?
Speech, language and communication difficulties (SLCDs)
Theoretical assumptions raise issues about links or dissociations among speech, language and communication disorders. The three terms are important in diagnosis and intervention, but tacit assumption is that impairment is rarely selective (all three are affected in one individual).
Inverted T-model: Sentences (pairings of sound and meaning) are produced by selecting elements from the lexicon, merging (spell-out) and then feeding the PF (SM) and LF (CI).
Objectives on ongoing research
- identify grammatical markers across atypical cognitive phenotypes which are consistently impaired,
- revisit deficient syntax in SLI, autism and Down’s syndrome
- compare standard and Cypriot Greek
Meta-analysis (Leivada 2015: Locus preservation hypothesis)
Collins (2005): morpho-syntactic difficulties associated with tense inflection appear across pathologies with different genetic causes (SLI, ASD, DS, Williams syndrome) => variable expressivity.
The effects of genetic disorders on FL appear to be surprisingly non-specific (affect same markers), but are also highly specific (target only some parts of language while consistently sparing others). May also appear in neurotypicals (eg effect of tiredness).
What is needed is a fine-grained analysis of locus of variation using distributed morphology framework (unlike inverted T model). One problem with previous research is its focus on English. This work is limited by the impoverished inflectional system of English, so it is interesting to include more richly inflected languages such as Greek. Research on these languages using quantitative methods suggests that what has been identified as impaired or deficient syntax in the literature is in fact something else.
Specific language impairment
delays in process of language acquisition. No inclusionary criteria, rather exclusionary, and heterogeneous disorder. Research in later 1990s reported particular characteristics in SLI Greek (omissions of definite article, but not indefinite; object clitics also suggested as vulnerable markers) but there is variation depending on data collection techniques and replication has not been successful.
What do omitted markers mean? Even if markers are omitted, their selectional requirements remain fully operative (Mastropavlou & Tsimpli, 2011). SLI research supports a delayed account rather than a deviant account, since a high percentage of correct use is ultimately manifested.
Phonetically conditioned differences in DS are due to the distinct physiology of the articular apparatus, coupled with a small residue of morphologically and phonologically conditional differences (Christodoulou, 2011). Final -s omission might suggest case problem but is in fact phonological. Grammatical markers are highly accurate (tense inflection, subject-verb agreement, pronouns, anaphors, tense and aspect). So impairment is at the level of external expression. English research does not agree, so perhaps English is not best language to study; more inflected languages are a better source.
ASD group (Terzi et al, 2012, 2014) had high performance in all types (pronouns, clitics), lowest performance in comprehension of clitics. Narrations are often underinformative, or lacking in temporal linking.
Variation across pathologies
Syntactic deficits are assumed on basis of omissions of free or bound markers. On closer analysis, the omissions boil down to of morphophonological markers or
Variation across pathologies can be formally captured by Locus Preservation Hypothesis
Through the Locus Preservation Hypothesis, a classical chicken and egg question can be addressed: Do poor conceptual resources and memory limitations result in an atypical grammar or does a grammatical breakdown lead to conceptual and memory limitations? Overall, certain morphological markers reveal themselves as highly susceptible to impairment, while syntactic operations are preserved, granting support to the first scenario. The origin of resilient syntax is explained from a phylogenetic perspective in connection to the “syntax-before-phonology” hypothesis.
Leivada, E., Papadopoulou, E., Kambanaros, M., & Grohmann, K. K. (2017).
In sum, syntactic operations are impenetrable to variation: their manifestations are universally preserved across the species. The computational system (“merge”) is intact, and problems occur only in externalisation, that is, at the phonetic interface or in terms of pragmatics
The high level of individual variation suggests that core syntax cannot be damaged or impaired, supported a “small UG” slimmed down framework, rather that a “big UG” modular framework. When surface syntactic deficits manifest, they do not follow from FL defects but stem from problems in the externalisation – consistent with small UG (rather than big UG). There is need for more cross-linguistic study, focusing on a spectrum of language pathologies taking perspectives drawing on comparative bilingualism, biolinguistics, and linguality.
- Use and interpretation of clitics or pronouns or articles depend on pragmatics. Grohmann expects deficiencies here to be apparent only in particular data collection circumstances, that is, only when contextual information is required.
- Neural footprint: granularity mismatch problem – current instruments do not allow linguistics to access fine level of granularity that interests them
- Does UG boil down to “merge”? Since animals can concatenate, does this mean they can merge (making UG not species-specific)? Animal concatenation is not considered to be an example of merge.