Lingustics 401, section 2HaleThe (Neuro-) Psychological Reality of MorphologySeptember 11, 2003Regularity: the speed of English verbsWilliam D. Marslen-Wilson and Lorraine K. Tyler. Dissociating types of mental computation. Na-ture, volme 382. June 1997.Marslen-Wilson and Tyler document the performance of brain-damaged patients in lexical-decisiontasks. One typ e of patient is primed by irregular (but not regular) verbs. Conversely, the othertyp e of patient is primed by regular (but not irregular) verbs.lexical decision deciding whether a stimulus is a real word or nonsensetimed lexical decision subject’s actual answer ignored and the time before response is recordedpriming a reduction in timed lexical decision latencyinterference an increase in timed lexical decision latencyNormal p e ople are known to be prime-able by related words spoken right before the stimulus to belexically-decided upon:relationship prime target priming?morphologicalregjumped jump YESmorphologicalirrfound find YESsemantic swan goose YESphonological gravy grave NOMarslen-Wilson and Tyler measured priming in patients afflicted with agrammatic aphasia. Suchpatients understand and are able to produce single words, but their comprehension and productionof grammatical combinations (including inflectional morphology) is spotty and inconsistent.patient quoted neurological assessmentJG, 66 1992 PET scan showed no active metabolism in left hemisphereDE, 41 1996 MRI scan revealed large left hemisphere lesionTS, 50 1996 CT scan reveal right hemisphere damage and changes tofrontal and occ ipital lobes. Patchy damage to left hemisphere.Marslen-Wilson and Tyler found a dissociation between regular and irregular past tenses in thesepatients.JG, DErelationship prime target priming?semantic swan goose YESmorphologicalregjumped jump NOmorphologicalirrfound find YESTSrelationship prime target priming?semantic swan goose NOmorphologicalregjumped jump YESmorphologicalirrfound find NO1Regularity: mistakes going out, but not coming inWilliam Badecker and Alfonso Caramazza. Morphological Composition in the Lexical Output Sys-tem. Cognitive Neuropsychology, volume 8, number 5. 1991.SJD is a 47-year-old college educated female who suffered a thromb oembolic left-hemispherestroke in June 1984...SJD’s speech is characterised by fluent, usually complete sentences with occasional mor-phological and function word errors, semantic paraphasias (e.g. brother for ‘sister’),phonemic paraphasias (e.g. cheri for ‘chair’), and hesitations for word-retrieval.use the word ‘darken’ in a sentenceIt gets darkly after eightThe day was darking be fore the storm• lexical decision: 99% correct for real words, 97.4% on nonwords• indicating which of a pair can be a person by pointing: 100%• reading individual words on cards: Adj 100%, N 98%, V 80%• one hundred English verbs: 60% on regulars, 92% on irregulars“Several of her responses were accompanied with a definition that was appropriate to the stimulus,but not to the response.”word on card what SJD saidcapitalize capsulation, you’re suppose d to use that with your namestarter startle-y, startled, like the car won’t work because of the ..,periodic peratid, like every week, supposed to have it re-...constantlylantern laenturd, you can have a candle or put some oil instead of electricityhostility hostages, like going to war, there something in betweenpublisher publish, someone who makes a magazine, book, newspaper−→ SJD understands, she just has trouble composing the morphemes in order to produce spokenoutput.2Inflection: kinds of fake Italian verbsAlfonso Caramazza, Alessandro Laudanna and Cristina Romani. Lexical access and inflectionalmorphology. Cognition, volume 28. 1988.Caramazza et al. (1988) used non-word interference results to argue for inflectionaldecomposition in Italian. They found that non-words composed of verbal stems withinappropriate inflections (e.g., eanteui, in which the first-conjugation verb root eant-occurs with the second-conjugation suffix -eui) took longer to rejec t in a lexical decisiontask than non-words composed of either real stems and nonce suffixes (e.g., eantoui) ornonce stems with real suffixes (e.g., eanzeui). In turn, these took longer to reject thannon-words composed entirely of nonce constituents (e.g., eanzoui).Chapter 9 of the Handb ook of Cognitive Neuropsychologynonword stimuliItalian verbs come in one of three conjugation types. The real way to say “you sing” in Italian is‘cantate.’cant -evi morphological legal nonwordto sing1st conjugationsecond person singular past,2nd conjugationcant -ovi no such affix, stem only is legalcanz -evi no such stem, affix only is legalcanzovi neither root nor affix legallexical decision resultsCANTEVI CANZEVI CANTOVI CANZOVItime in ms 875 809 781 760percent error 21.0 7.7 11.8 5.8Italian speakers take more time and make more mistakes deciding if malformed morphologicalcombinations are words or not.3Compounding: making the problem worseWilliam Badecker. Lexical composition and the production of compounds: evidence from errors innaming. Language and Cognitive Processes, volume 16, number 4. 2001.CSS [65 years] suffered a left cerebro-vascular accident in May 1990 resulting in lexi-cal impairments manifested in reading, repetition, and oral and written naming tas ks.Apart from his word-finding difficulties, CSS’s spoken output is fluent and he exhibitsno apparent comprehension impairment.picture-naming“butterfly”naming-to-definitionSlender-bodie d insect with broad,often brightly colored wings.“butterfly”When the stimuli have just one morpheme, CSS is about 80% c orrectwhat CSS sees what CSS sayssemantic errorscrab lobsterglobe atlasphonological errorspenguin pendulumwrist wrisp [sic]When the stimuli have two morphemes, CSS is only about 50% correct!what CSS sees what CSS sayscomponent errorssundial sunphonological neol ogismdragonfly doctor flydoll house toll houseweight lifter weight loafersemantic neologismashtray ash vasewheelchair wheel pillcheer leaders gym leadersordering errorstrash can can trashsnow shoe shoe snowCompound neologisms involved same kinds of errors as he makes with monomorphemic words. Thissuggests that CSS’s ability to deal with morphological componding has been selectively