10.1016@j.jneuroling.2012.07.003 | Lexicon | Verb

Please download to get full document.

View again

of 18
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information Report
Category:

Documents

Published:

Views: 4 | Pages: 18

Extension: PDF | Download: 0

Share
Related documents
Description
journal of neurolinguistics
Transcript
  Does verb type affect action naming in speci 󿬁 c languageimpairment (SLI)? Evidence from instrumentality and namerelation Maria Kambanaros * University of Cyprus, Cyprus Acquisition Team, 75 Kallipoleos, P.O. Box 20537, 1678 Nicosia, Cyprus a r t i c l e i n f o  Article history: Received 21 June 2012Received in revised form 13 July 2012Accepted 17 July 2012 Keywords: Instrumental verbsNon-instrumental verbsGreekDiglossiaLemmaLexemeSemantic complexity a b s t r a c t Children with speci 󿬁 c language impairment (SLI) have particularproblems using verbs. However, not much is known about howchildren with SLI retrieve different  types  of verbs. In the presentstudy, bilectal Greek-speaking children with and without SLI wereassessed on naming of different verbs held constant for argumentstructure but manipulated for lexical-semantic and phonological-lexical features.School-aged children with SLI as well as typically developing age-and vocabulary-matched peers named 39 colored photographsrepresenting actions in a confrontation naming task. Stimuliincluded actions involving an instrument (e.g., sweeping), i.e.instrumental verbs, and actions that have a name-relationwith theinstrument (e.g., sawing), i.e. name-related instrumental verbs aswell as actions not involving an instrument (e.g., climbing), i.e.non-instrumental verbs.Instrumental verbs were signi 󿬁 cantly more dif  󿬁 cult to retrievethan non-instrumental verbs for children with SLI and typicallylanguage-developing controls. In contrast, instrumental verbs witha name relation to the noun were signi 󿬁 cantly easier to name thaninstrumental verbs without a name relation for all groups.Children with SLI performed on par with vocabulary-matchedpeers. The results based on error types suggest that the greaterdif  󿬁 culties children with SLI have with action naming is lexical-semantic in nature. *  Corresponding author. Tel.:  þ 357 22 89 51 94; fax:  þ 357 22 75 03 10. E-mail address:  kambanaros@gmail.com. Contents lists available at SciVerse ScienceDirect  Journal of Neurolinguistics journal homepage: www.elsevier.com/locate/ jneuroling 0911-6044/$  –  see front matter    2012 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.jneuroling.2012.07.003  Journal of Neurolinguistics 26 (2013) 160 – 177  The  󿬁 ndings indicate a need to examine the link between the verbnaming de 󿬁 cit in SLI to structural and functional abnormalities inBroca ’ s area.   2012 Elsevier Ltd. All rights reserved. 1. Introduction Children presenting with speci 󿬁 c language impairment (SLI) have variable de 󿬁 cits in differentcomponents of the grammar (syntax, morphology, phonology) as well as other aspects of language(e.g.,vocabulary)intheabsenceoffactorsthattypicallyaccompanylanguageproblemssuchashearingimpairment,lownon-verbalIQ,neurologicaldamage,orsocio-emotionaldeprivation.SLIisconsidereda neurodevelopmental disorder in which complex genetic and multiple environmental risk factors areimplicated (see Bishop, 2006).One recent cross-linguistic characteristic to emerge is a de 󿬁 cit in grammatical word class pro-cessing, with verbs typically more impaired compared to nouns (Kambanaros, Grohmann, &Theodorou, 2010; Kambanaros, Psahoulia, & Mataragka, 2010; Sheng & McGregor, 2010b). Indeed, the verb lexicon of SLI children has generated much research over the last thirty years concerningin 󿬂 ectional morphology (see Rice, 2009), argument-structure complexity (see Pizzioli & Schelstraete, 2008), and lexical-semantic access (see Kambanaros & Grohmann, 2010, 2011). The latter is a new domainandtheleastresearched(ofthethree)butthetopicofthepresentstudy.Itrelatestodif  󿬁 cultieswith verb processing children with SLI have outside the classic area of morphosyntax, but insteadwithin the semantic and/or phonological lexicon, that is, in accessing and retrieving stored verbmeanings from memory.At the neural level, verb processing is supported by the left frontal cortex, mainly Broca ’ s area(Shapiro, Moo, & Caramazza, 2006). Most recently, research using functional imaging (voxel-basedmorphometry) identi 󿬁 ed structural and functional abnormalities in the left inferior frontal gyrus orBroca ’ s area in children and adolescents with SLI (Badcock, Bishop, Hardiman, Barry, & Watkins, 2011).The focus of the present research is on verbs that require an obligatory instrument to perform anaction (e.g.,  sweep  requires a broom) versus verbs that do not (e.g.,  climb  a mountain/tree/ladder), andbetween verbs that have a lexical link such as a name relation between the instrument and the verb(e.g.,  the saw  –  to saw ), which to date have only been reported in acquired language breakdown (seeKambanaros, 2009b, and references within).Every verb meaning has two key components. First, they involve an event schema representing anevent type which can have simple structural meaning of the type [x ACT < MANNER  > ] or a morecomplex one such as [[x ACT < MANNER  > ] CAUSE [BECOME [y  < RES-STATE > ]]]. Second, a root ischaracterized by the ontological type chosen from a  󿬁 xed set of options (e.g., state, result, location,manner, etc.) expressing the idiosyncratic properties of the verb meaning. For example, if an object ornoun(N)namesaninstrument,thecorrespondingactionorverb(V)means “ usethatinstrumentforitspurpose ” : instrument / [x ACT  < INSTRUMENT > ] (e.g.,  brush ,  hammer  ,  saw ,  shovel ,  sweep ). Associa-tions are probably not linguistic, but rather re 󿬂 ect general cognitive principles. Instrument roots arereally a subtype of manner roots, behaving like them in all respects. In addition, instrument roots areintegrated into schemas as modi 󿬁 ers of predicates, and as such instrumental verbs are by de 󿬁 nitionsemantically more complex (see Rappaport Hovav & Levin, 2010, on all points regarding verbmeanings).In Indo-European languages like Modern Greek, the language under investigation (as well asEnglish, Dutch, French, etc.), instrumentality can be expressed by means of a prepositional phrase toverbalize the instrument (e.g.,  The man is sawing the wood with a saw 1 ) or simply by the instrumental 1 The instrument is rarely used with the instrumental verb in the same sentence, however, as this gives it a semantically butnot necessarily syntactically unusual reading ( Jonkers, 1998). M. Kambanaros / Journal of Neurolinguistics 26 (2013) 160 – 177   161  verb itself (e.g.,  The man is sawing  ). Instrumental verbs can be divided into two types: (i) instrumentalverbs with a name relation to the instrument (e.g.,  to saw  and  the saw ), where one and the same(phonological)wordformrepresentsanounaswellasaverb,and(ii)instrumentalverbswithnonamerelation to the instrument (e.g.,  to sew  and  needle ).Inastem-basedlanguagelikeGreek, name-relatedN – Vpairsarenotrepresentedbyonewordformbut instead share the same stem or root. Speakers of Greek need to correctly in 󿬂 ect the stem witheither a nominal or a verbal in 󿬂 ection to produce the instrument/noun and instrumental verb,respectively. Forexample,  prion –  is the stem foreither  ‘ (a) saw ’  or  ‘ (to) saw ’ . It has little semantic valueatthestagepriortoin 󿬂 ection.Whenin 󿬂 ectedwiththeaf  󿬁 x – i itbecomesthenominativeoraccusativecase-marked singular noun lemma  prioni  ( ‘ (a) saw ’ ) and with the af  󿬁 xes  – iz  þ – o  it becomes the  󿬁 rstperson singular present tense indicative verb lemma  prionizo  ( ‘ (I) saw ’ ).Based on models of the lexicon such as Levelt ’ s (1989, 2001), the conceptual-semantic and the syntactic information of a lexical entry (i.e. the lemma-level information) is needed for grammaticalencoding. In response to an action picture, a verb lemma is activated, specifying lexical-syntacticinformation such as the verb ’ s argument structure, tense, person, and number information. Forexample, the verb  sew  has two semantic arguments, an external argument (the agent role) and aninternal argument (the theme or patient role); syntactically, we say that as a transitive verb, it takesone object to project a verb phrase and then combines with a subject. In addition, at the level of thelemma, lexical-semantic knowledge about the verb is activated and may include speci 󿬁 c attributessuch as [  instrument]. Speci 󿬁 cally, for the verb  sew  the objects  needle  and  thread  may also be part of the(pictured)actionsemantics.Hence,instrumentalityisafeatureoperatingatthelevelof thelemma.At the second stage of word retrieval, the lexeme or word form corresponding to the selectedlemma is phonologically speci 󿬁 ed. Lexemes contain information about the phonology of a word(number of syllables, prosody, segmentation) and its morphology (verb in 󿬂 ections). Lexeme-levelinformation for  saw  includes the phonemes /s/ and / ɔ :/, and the morphophonological variant thatallows it to become a verb (or a noun) in Modern Greek. Hence, name relation is a phonologicalproperty operating at the level of the word form. For successful lexicalization in the discrete two-stepmodel of  Levelt (1989), adequate amounts of informationwithin both the lemma and the lexeme mustbe available and activated for the target item to be retrieved. This presupposes an effective  󿬂 ow of unidirectional information about the lexical entry between these two levels of processing.Lexical entries are no longer considered to be listed as words organized in a single-word formatwithin a (mental) dictionary. Instead they are integrated and interconnected based on shared infor-mationattheconceptual,syntactic,morphological,and/orphonologicallevels, thusallowingforeasieraccess and retrieval. Speci 󿬁 cally, Levelt (1989) describes this relationship between lexical items withsimilar lemma and/or lexeme information as  intrinsic  . To illustrate, the verbs  sweep  and  mop  areintrinsicallyrelatedbecausetheybothareactionsofcleaning(manner)involvinganitemwithapoleorhandle (instrument). This of course results in multiple simultaneous activations of lexical cohorts ascompetitors. Yet, once all available information about the speci 󿬁 c target word is integrated and pro-cessed, only one remains active, and all other semantic and/or phonological competitors are deacti-vated so that the target word is eventually produced. Furthermore, verbs with additional and morespeci 󿬁 csemantic components(e.g.,compare sweep with clean )are consideredmorecomplex(Breedin,Saffran, & Schwartz, 1998). 2. Background to the study  Children with SLI have particular problems learning and using verbs, and this characteristic isconsidered, at least for English, a clinical marker of SLI (Rice, 1991, 2003). A small body of literature, usually based on the analysis of spontaneous language, has shown that children with SLI usesemantically less speci 󿬁 c verbs, compared to typically developing peers. They are also reported to relymoreheavilyon a subsetofnon-speci 󿬁 c verbstodescribe events and actions oftentermed general-all-purpose (GAP) verbs, such as  do ,  get  ,  have ,  make ,  put  ,  come ,  give ,  look ,  play ,  see ,  take , or  want   (Conti-Ramsden & Jones, 1997; de Jong, 1999; Kelly, 1997; Rice & Bode, 1993; Watkins, Rice, & Moltz, 1993) d and their counterparts across languages (e.g., Stavrakaki, 2000, for Modern Greek). In allcases, the lexical-semantic speci 󿬁 cation of the verb is so general that it can be used in a multitude of  M. Kambanaros / Journal of Neurolinguistics 26 (2013) 160 – 177  162  contexts. The use of GAP verbs as stand-ins for more speci 󿬁 c verbs has been linked to dif  󿬁 cultieschildren with SLI have with fast mapping of verb meanings or world-to-word mapping of the action(Rice & Bode, 1993; Stavrakaki, 2000), extracting verb meanings or the greater polysemy of verbs (de  Jong,1999), and establishing categorical and/or semantic boundaries for verbs (Kelly, 1997). In the study by de Jong (1999), lexical access for speci 󿬁 c verb types was targeted for investigationusing a confrontation naming task. By means of video clips, four sets of verbs differing in mannerfeatures were investigated: going verbs (e.g., skipping), putting verbs (e.g., sewing), cutting verbs(e.g., cutting), and cleaning verbs (e.g., mopping). They were all shown to participating Dutch school-aged children (mean age 7; 8 years) with and without SLI who were required to name the actionappearing on the screen. For the purpose of the present study the categoryof   “ going ”  verbs will not bediscussed further. Of the seventeen verbs belonging to the other three categories, eight were actionsthatinvolvedaninstrumentaspartoftheiractionsemantics(e.g., sew , stick/glue ,  pin , cut  , lawn mowing  , scrub/mop , and  hoover  ). 2 Children with SLI committed mainly substitution errors by either replacingthe semantically-speci 󿬁 c action name with a general-all-purpose (GAP) verb construction (e.g., for mop/scrub / ‘ make clean ’ ) or omitting the instrument feature but retaining the manner of the action(e.g., for  pin / ‘ hang ’ ). de Jong (1999, p. 178) claimed that children with SLI adopt a  “ lexical under-speci 󿬁 cation strategy ”  and as such have dif  󿬁 culty matching an action with the proper verb.Current literature around lexical processing in SLI lends support to this  󿬁 nding that children withSLI may have dif  󿬁 culty with lexical-semantic processes such as lexical selection possibly due to poorlyde 󿬁 ned and badly organized semantic representations at the lemma level (Alt & Plante, 2006;Brackenbury&Pye,2005;Lahey&Edwards,1999;McGregor,Newman,Reilly,&Capone,2002;Sheng& McGregor, 2010a). On the other hand, strong evidence exists for an account attributing lexical pro-cessing de 󿬁 cits to weak underlying phonological representations in the lexicon and limitedphonological-lexicalprocessingskillsinchildrenwithSLI(Chiat,2001;Seiger-Gardner&Brooks,2008; Velez & Schwartz, 2010). Lastly, both semantic and phonological de 󿬁 cits are equally implicated in thenature of word-learning de 󿬁 cits in SLI (Kambanaros, Grohmann, et al., 2010; Kambanaros, Psahoulia, et al., 2010; Nash & Donaldson, 2005). To the best of my knowledge, the present study is the  󿬁 rst to focus on the retrieval of instrumentalverbs in children with a developmental language disorder such as SLI. The impetus for the researchstems from the fact that developmental lexical retrieval de 󿬁 cits in SLI children appear very similar inform to acquired lexical retrieval de 󿬁 cits in adults (see Friedmann, Biran, & Dotan, 2012, for a  󿬁 rstexplanation of developmental anomia and its subtypes).Instrumentality, a conceptual-semantic factor, has been shown to signi 󿬁 cantly affect verb retrievaleither in a positive way by facilitating action naming in  󿬂 uent aphasia (Bastiaanse & Jonkers, 1998; Jonkers & Bastiaanse, 1996, 2007; Kambanaros, 2009a; Kambanaros & van Steenbrugge, 2006) or in anegativewaybyhinderingverbretrievalinnon- 󿬂 uentaphasia(Bastiaanse&Jonkers,1998; Jonkers& Bastiaanse,1996,2007) d orto have noeffectonactionnamingirrespective ofaphasiatype (Kemmerer& Tranel, 2000). Similarly, name relation, a phonological-lexical feature operating at the level of thephonological word form, has revealed contradictory results on action naming in individuals withaphasia: a positive effect on instrumental verbs with a name relation easier to retrieve than thosewithout ( Jonkers & Bastiaanse, 1996, 2007; Kemmerer & Tranel, 2000), a negative effect (Bastiaanse, 1991; Kambanaros & van Steenbrugge, 2006), and no effect (Bastiaanse, 1991; Jonkers, 1998; Kambanaros & van Steenbrugge, 2006). 3 Given the evidence from the SLI literature so far, the hypothesis was that both instrumental verbs(semantic complex) and name-related instrumental verbs (semantic  þ  phonological information)should be more dif  󿬁 cult to retrieve on a confrontation naming task for children with SLI compared tonon-instrumental verbs and non-name-related instrumental verbs, respectively. Children with SLI are 2 There were two examples of the verb  cut  , one with scissors and the other with a knife. The verb  pin  incorporateda thumbtack. 3 In Bastiaanse (1991), two anomic patients were described: one demonstrated a negative effect of name relation on actionnaming, whereas the other showed no effect whatsoever. In Kambanaros and van Steenbrugge(2006), a negative effect of namerelation was reported in L2 (English) and no effect in L1 (Greek) for 12 bilingual patients with anomic aphasia. M. Kambanaros / Journal of Neurolinguistics 26 (2013) 160 – 177   163
Recommended
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks