A Neural Dissociation Within Language
Total Page:16
File Type:pdf, Size:1020Kb
A Neural Dissociation Within Language: Evidence that the Mental Dictionary is Part of Declarative Memory, and that Grammatical Rules are Processed by the Procedural System The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters Citation Ullman, Michael T., Suzanne Corkin, Marie Coppola, Gregory Hickok, John H. Growdon, Walter J. Koroshetz, and Steven Pinker. 1997. A neural dissociation within language: Evidence that the mental dictionary is part of declarative memory, and that grammatical rules are processed by the procedural system. Journal of Cognitive Neuroscience 9(2): 266-276. Published Version doi:10.1162/jocn.1997.9.2.266 Citable link http://nrs.harvard.edu/urn-3:HUL.InstRepos:3600798 Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of- use#LAA A Neural Dissociation within Language: Evidence that the Mental Dictionary Is Part of Declarative Memory, and that Grammatical Rules Are Processed by the Procedural System Michael T. Ullman Institute for Cognitive and Computational Science Georgetown University Suzanne Corkin MIT Marie Coppola University of Rochester Gregory Hickok University of California, Irvine John H. Growdon and Walter J. Koroshetz Massachusetts General Hospital Steven Pinker MIT Abstract W Language comprises a lexicon for storing words and a gram- memory impairment in Alzheimer’s disease, led to more errors mar for generating rule-governed forms. Evidence is presented with irregular than regular and novel verbs. Grammatical that the lexicon is part of a temporal-parietalhnedial-temporal difficulties in anterior aphasia, and the general impairment of “declarative memory” system and that granlmatical rules are procedures in Parkinson’s disease, led to the opposite pattern. processed by a frontamasal-ganglia “procedural” system. Pa- In contrast to the Parkinson’s patients, who showed sup tients produced past tenses of regular and novel verbs (looked pressed motor activity and rule use, Huntington’s disease pa- and plagged), which require an -ed-suffixation rule, and irregu- tients showed excess motor activity and rule use, underscoring lar verbs (dug), which are retrieved from memory. Word-finding a role for the basal ganglia in grammatical processing. W difficulties in posterior aphasia, and the general declarative INTRODUCTION frontal cortex, including Broca’s area, and links between lexical memory and left temporal and parietal cortex Two capacities give human language its vast expressive (Geschwind, 1965;Goodglass, 1993;Wernicke, 1874).The power. One is a “mental lexicon” containing thousands speech of aphasics with frontal (anterior) lesions is often of words, each a memorized, arbitrary sound-meaning “agrammatic,”with a breakdown of sentence structure pairing. The other is a “mental grammar” of generative and the omission or misuse of grammatical morphemes, rules that combine words into an infinite number of while access to content words such as verbs and nouns larger words, phrases, and sentences (Chomsky, 1965; De is less dramatically impaired (Damasio, 1992;Goodglass, Saussure, 1959;Pinker, 1994). Previous evidence has sug- 1993). Electrophysiological and functional neuroimaging gested links between grammatical processing and left studies have also implicated frontal regions in grammati- 0 1997 Massachusetts Institute of Technology Journal of Cognitive Neuroscience 9:2,pp. 266-276 cal processing (Kluender & Kutas, 1993; Neville et al., matical processing (Lieberman et al., 1992), performing 1991; Stromswold et al., 1996). Aphasics with temporal operations comparable to those done for motor pro- or parietal (posterior) lesions often speak in relatively gramming. Damasio and Damasio (1 992), noting the ana- intact sentence structures, but have severely impaired tomical interconnections between cortical language access to content words (Damasio, 1992; Goodglass, areas and the basal ganglia, predict that “the basal ganglia 1993). The dissociation, however, is imperfect and still serve to assemble the components of complex motions controversial (Bates & Wulfeck, 1989; Goodglass, 1993). into a smooth whole, and it seems reasonable that they Most groups of aphasics tested have complex lesion might perform an analogous function in assembling patterns, making symptom-lesion correlations imperfect. word-forms into sentences.” Moreover, testing the dissociation has been problematic We tested this memoryhule dissociation by devising because tasks probing for grammar and for memory have a task based on a simple linguistic system in which differed in ways other than their use of the two capaci- reliance on grammar and lexicon differs, while other ties. In this study we confirm the dissociation using a factors are held constant. Regular (look-looked) and ir- simple language task in which the use of the two Wstic regular (dig-dug) past tense forms of verbs are well- capacities is contrasted while other factors, such as com- matched in complexity (one word), syntax (tensed), and plexity, meaning and task demands, are held constant. meaning (past). But regular verbs are predictable in form We also relate the dissociation to larger principles of (verb stem + e9,and new ones are constantly being neural organization. Cutting across the division of the added (tir?ed, mosbed), whereas irregular verbs are un- brain into systems such as language and visual percep predictable (compare sing-sang, jling$ung, bring- tion, there is an orthogonal division into two major kinds brought), and constitute a fixed list. A simple theory is of memory systems (Cohen & Squire, 1980; Mishkin et that irregular forms are memorized, and regular forms al., 1984; Damasio & Damasio, 1992; Squire et al., 1993). are generated by a rule. The rule comprises two opera- One is a declarative memory system underlying the tions: copying the stem, and adding a suffix. Regulars and learning and storage of information about facts and irregulars interact as follows: Retrieval of an irregular events. It is subserved by a medial temporal circuit con- blocks the rule (dug preempts digged);when an irregu- nected largely with neocortical areas in the temporal lar is not successfully retrieved, the rule may be applied, and parietal lobes, with the medial temporal compo- resulting in “overregularization”errors such as digged. nents consolidating memories that are eventually stored Alternative theories have been proposed. In one, regu- in neocortex (Cohen & Squire, 1980;Mishkin et al., 1984; lars and irregulars are both computed by rules (Chom- Squire et al., 1993; Suzuki & Amaral, 1994). The other is sky & Halle, 1968; Halle & Mohanon, 1985), with aprocedural memory system for the learning and proc- memory compressed to the minimum information nec- essing of motor, perceptual, and cognitive skills. It is essary. In another, regulars and irregulars are both com- subserved by basal ganglia circuits connected largely puted by a connectionist associative memory, with all with frontal cortex (Cohen & Squire, 1980; Gabrieli et rules eliminated (Rumelhart & McClelland, 1986; al., 1993; Heindel et al., 1989; Mishkin et al., 1984; Saint- MacWhinney & Leinbach, 1991). Recent evidence from Cyr et al., 1988; Squire et al., 1993). These parallel basal the structure, processing, and acquisition of several lan- ganglia circuits are functionally segregated;each receives guages has bolstered the memory/rule theory (Marcus projections from particular ipsilateral cortical and sub- et al., 1992; Marcus et al., 1995; Pinker, 1991; Pinker & cortical areas, and projects via the thalamus to a particu- Prince, 1988; Prasada & Pinker, 1993), although contro- lar ipsilateral frontal lobe area. Thus a “motor circuit” versy persists, with some arguing that either an associa- projects to frontal motor areas, while other circuits proj- tive memory, or a set of rules, suffices (Hare & Elman, ect to other frontal areas. The different circuits have 1995; Ling & Marinov, 1993; Plunkett & Marchman, similar synaptic organizations within the basal ganglia 1993). It is therefore important to confirm the mem- (Middleton & Strick, 1994; Alexander et al., 1990;Young ory/rule distinction in the form of a neural dissociation, tk Penney, 1993). and to use it to illuminate the functions of the major Given that word forms are like facts in being arbitrary, brain systems underlying language and their relation to and in possibly having storage sites in temporal and overall brain organization. parietal regions, the temporal-parietavmedial-temporal If indeed (a) irregulars are stored words, (b) the tem- declarative memory system may subserve words as well poral-parietal/medial-temporaldeclarative system under- as facts and events. Given that rules are like skills in lies word memory, (c) regulars are rule-products,and (d) requiring the coordination of procedures in real time, the frontal/basal-ganglia procedural system underlies the and in possibly having neural loci in frontal regions, the processing of rules, then the following double dissocia- frontal/basal-ganglia procedural system may process tions are predicted. Patients with impairments of lexical grammatical rules as well as motor and perceptual skills. memory, or more generally, of declarative memory, from Basal gangha circuitry may project to Broca’s area damage to temporal or parietal