EFFECTS OF GRAMMATICAL PRIMING IN WORD NAMING ARE SUBSTANTIAL AND ROBUST, AT LEAST IN SERBIAN Dejan Lalović Department of Psychology, Faculty of Philosophy. - презентация

1 EFFECTS OF GRAMMATICAL PRIMING IN WORD NAMING ARE SUBSTANTIAL AND ROBUST, AT LEAST IN SERBIAN Dejan Lalović Department of Psychology, Faculty of Philosophy University of Belgrade Republic of Serbia

2 EFFECTS OF GRAMMATICAL PRIMING IN WORD NAMING ARE SUBSTANTIAL AND ROBUST, AT LEAST IN SERBIAN DEFINING TERMS

3 SEMANTIC PRIMING PRESENTATION TALK ASSOCIATIVE PRIMING BOREDOM TALK GRAMMATICAL PRIMING THIS TALK primetarget

4 ! P RIMED ONLINE WORD NAMING ( READING ALOUD ) TASK THROUGH TUNNEL voice key recorded online RT + errors coded Read the first word in each pair silently Read the following word in each pair aloud (P RIME ) (T ARGET )

5 EFFECTS OF GRAMMATICAL PRIMING IN WORD NAMING ARE SUBSTANTIAL AND ROBUST?, AT LEAST IN SERBIAN THEORETICAL BACKGROUND, AIMS OF THE STUDY

6 L ITERATURE ON GRAMMATICAL PRIMING EFFECTS ON VISUAL WORD PROCESSING IS EQUIVOCAL : G RAMMATICAL PRIMING CONSISTENTLY YIELDS EFFECTS IN LEXICAL DECISION TASK BOTH IN E NGLISH ( E. G. G OODMAN ET AL. (1981), S EIDENBERG ET AL. (1984), W RIGHT & G ARRET (1984), W EST & S TANOVICH (1986)), AND S ERBIAN ( E. G. L UKATELA ET AL. (1982), K ATZ ET AL. (1994)). BUT T HESE EFFECTS ARE DIMINISHED WHEN NAMING TASK IS EMPLOYED ( E. G. S EIDENBERG ET AL., (1984) IN E NGLISH, C ARELLO ET AL., (1988) IN S ERBIAN )! PREVAILING VIEW: G RAMMATICAL PRIMING INFLUENCES STRATEGIC PROCESSES, THAT OCCUR UPON THE WORD HAS BEEN RECOGNIZED, WITHOUT INFLUENCING WORD RECOGNITION PER SE.

7 HOWEVER, ROBUST EFFECTS OF GRAMMATICAL PRIMING IN WORD NAMING HAVE BEEN OBSERVED IN AT LEAST TWO STUDIES : * S UCH SYNTACTICAL PRIMING EFFECT WAS EQUAL TO, OR OF COMAPRABLE MAGNITUDE IN BOTH WORD PROCESSING TASKS ! W EST & S TANOVICH (1986) EMPLOYED SENTENCE FRAGMENTS AS SYNTACTICAL PRIMES TO CONSTRAIN THE GRAMMATICAL CLASS OF THE TARGET WORDS IN BOTH LEXICAL DECISION AND NAMING TASK. * G RAMMATICAL PRIMING EQUALLY FACILITATED E NGLISH INFLECTED AND UNIFLECTED NOUNS AND VERBS, REGARDLESS OF SOME EXPERIMENTAL MANIPULATIONS ( TO BE EXPLAINED LATER ON ). B OWEY (1996) EMPLOYED SINGLE - WORD PRIMES ( DEMONSTRATIVE PRONOUNS AND AUXILIARY VERBS ) TO CONSTRAIN THE TARGET WORD CLASS ( SUBJECTS FOURTH - GRADE CHILDREN ).

8 C LAIMS THAT GRAMMATICAL PRIMING IS LIMITED TO LEXICAL DECISION TASK IN E NGLISH APPARENTLY SEEM PREMATURE ! THE ARGUMENT: I F GRAMMATICAL CONTEXT PLAYS SOME ROLE IN E NGLISH WORD RECOGNITION THEN, BY INTUITION, ITS ROLE IN S ERBIAN WORD RECOGNITION SHOULD BE MUCH MORE PROMINENT. T HE INTUITION RESTS UPON CERTAIN MAJOR PROPERTIES OF S ERBIAN GRAMMAR

9 S ERBIAN MORPHOLOGY IS, UNLIKE E NGLISH, HEAVILY INFLECTED. G RAMMATICAL ATTRIBUTES OF OPEN CLASS S ERBIAN WORDS ARE SPECIFIED BY INFLECTIONAL SUFFIXES. E. G. masculine noun forms and cases formcase(s) prstnom. s. prstagen. s, acc. s, g. pl. prstudat. s, loc. s. prstominst. s. prstinom. pl. prsteacc. pl. prstimadat. pl, loc. pl, inst. pl. prst = finger + 6 feminine noun forms, + 7 neuter noun forms…

10 S ERBIAN SYNTAX IS A RATHER FREE - ORDER ONE : Basically: Is this a way to earn own house, enjoying perch each evening!? Zar se tako teče kuća svako veče pored smuđa!? Zar se tako kuća teče, pored smuđa svako veče!? Zar se kuća teče tako pored smuđa veče svako!? (Laza Kostić, famous Serbian romantic poet)

11 Зар се кућа тече тако, поред смуђа вече свако!? Поред смуђа вече свако, кућа зар се тече тако!? Zar se kuća teče tako, pored smuđa veče svako!? Pored smuđa veče svako, kuća zar se teče tako!? subject, nominativ case Зар ћеш кућу стећи тако, поред смуђа вече свако!? Поред смуђа вече свако, кућу зар ћеш стећи тако!? Zar ćeš kuću steći tako, pored smuđa veče svako!? Pored smuđa veče svako, kuću zar ćeš steći tako!? object, accusative case ******************************* Kućom zar se diči neko ko je za kafanskim stolom steko!? Zar se kućom diči neko ko je za kafanskim stolom steko!? object, instrumental case L AZA K OSTIĆ S VERSES EVOLUTION INTO THE BEĆARAC

12 A LEXICAL ( OPEN - WORD CLASS ) MORPHOLOGICAL AND THUS SYNTACTICAL AMBIGUITY IN S ERBIAN IS EASILY RESOLVED BY MODIFIERS – CLOSE - CLASS FUNCTION WORDS ( E. G. PREPOSITIONS, PRONOUNS ) AND MORPHOLOGICALLY CONGRUENT ADJECTIVES PRECEEDING THE WORD BEING MODIFIED. G RAMMATICAL AGREEMENT IN S ERBIAN THEREFORE SHOULD BE A POWERFUL TOOL FOR SPECIFYING WORD S THEMATIC ROLE IN SENTENCE.

13 C RITICAL FEATURES OF THE STUDY PRESENTED T ARGET WORDS OF TWO CLASSES WERE PRIMED INSTEAD OF PRIMING GRAMMATICAL FORMS WITHIN A SINGLE WORD CLASS. S INGLE - WORD PRIMES THAT UNEQUIVOCALLY CONSTRAINED TARGET WORD CLASS WERE EMPLOYED IN THE CONSTRAINED EXPERIMENTAL CONDITION. N EUTRAL LINGUISTIC PRIMES WERE EMPLOYED IN THE UNCONSTRAINED EXPERIMENTAL CONDITION. 1. T O RE - INVESTIGATE EFFECTS OF GRAMMATICAL PRIMING ( I. E. GRAMMATICAL CONGRUENCY ) ON S ERBIAN WORD NAMING A IMS OF THE PRESENT STUDY :

14 TYPICAL PRIME -TARGET PAIRS AS A FUNCTION OF WORD TYPE AND CONSTRAINT CONSTRAINEDUNCONSTRAINED PRIMESTARGETSPRIMESTARGETS ZBOG/BEZ (BECAUSE/WITHOUT) BLUZE/TARABE (BLOUSE/FENCE) I/ILI (AND/OR) GREDE/PALICE (JOIST/BAT) NOUNS KROZ/UZ (THROUGH/UP) SCENU/TERASU (SCENE/TERACE) I/ILI (AND/OR) GLINU/KOLIBU (CLAY/COTTAGE) ON/ONA (HE/SHE) SPASE/RUKUJE (SAVES/HANDLES) I/ILI (AND/OR) GREBE/ZABODE (SCRAPES/STICKS) VERBS ONI/ONE (THEY, MASC./FEM.) SNUJU/DIRAJU (DREAM/TOUCH) I/ILI (AND/OR) BRINU/DOVEDU (CARE/BRING)

15 E XPERIMENTAL MATERIALS, SIMPLIFIED TARGETS tears night hill fog steals cries love trust VERBS NOUNS In grammatically neutral (unconstrained) condition, both noun and verb targets were primed by conjectures AND/OR, which did not preclude neither w. class nor form. PRIMES PERSONAL PRONOUNS she he they m. they f. without because up through PREPOSITIONS and/or

16 AIMS OF THE PRESENT STUDY: 2. TO INVESTIGATE POSSIBLE INFLUENCE OF SOME MAJOR EXPERIMENTAL MANIPULATIONS ON GRAMMATICAL PRIMING EFFECTS. M AJOR EXPERIMENTAL MANIPULATION = S TIMULUS O NSET A SYNCHRONY ! THROUGH TUNNEL PRIME EXPOSITION ISIO NLINE REACTION TIME SOA AND,

17 ! D ELAYED NAMING TASK – ONE WAY TO CONTROL FOR THE NON - COGNITIVE FACTORS IN ( ONLINE ) WORD NAMING : TUNNEL voice key recorded delayed RT Read the words silently as they appear on the screen, but pronounce them aloud only after the brackets appear! ( )

18 EFFECTS OF GRAMMATICAL PRIMING IN WORD NAMING ARE SUBSTANTIAL AND ROBUST, AT LEAST IN SERBIAN RESULTS

19 RT [ms] 600 ms 250 ms Unconstr. Constr. Unconstr. O NLINE NAMING Verbs-Nouns p<.001 SOA p<0.05 SOA x V-N n. s. SOA x Constraint n. s. Constraint p< ms 19 ms SOA ERR% 600 ms250 ms Constr. Unconstr. Constr. SOA p<0.05 Constraint n. s. (p=0.09) Verbs-Nouns p<.01 SOA x Constraint n. s. SOA x V-N n. s. SOA

20 D ELAYED NAMING Targets in Con. Targets in Uncon. N O MAIN EFFECTS, NO INTERACTIONS Targets in Con. Targets in Uncon. RT [ms] Constr. 600 ms250 ms Unconstr. SOA O NLINE - D ELAYED NAMING RT [ms] 600 ms250 ms SOA p<0.05 Constraint p<.001 Verbs-Nouns p<.001 SOA x Constraint n. s. (27 vs. 19 ms) SOA x V-N n. s. SOA 125 ms 183 ms BUT! SOA

21 EFFECTS OF GRAMMATICAL PRIMING IN WORD NAMING ARE SUBSTANTIAL AND ROBUST, AT LEAST IN SERBIAN SUMMARY, CONCLUSIONS AND SOME OPEN QUESTIONS

22 C LAIMS THAT GRAMMATICAL PRIMING IS RESTRICTED TO LEXICAL DECISION TASK CAN BE REFUTED ON EMPIRICAL GROUNDS. F URTHER, G RAMMATICAL PRIMING EFFECTS ARE NOT LIKELY TO BE INFLUENCED BY SOME MAJOR STRATEGIC PROCESSES DIRECTLY ATTRIBUTABLE TO THE SOA INTERVAL MANIPULATIONS. S UBSTANTIAL FACILITATIVE EFFECTS OF GRAMMATICAL PRIMING ON S ERBIAN WORD NAMING ( ABOUT TWICE AS LARGE IN MAGNITUDE AS THOSE OBTAINED IN E NGLISH ) WERE OBSERVED IN ADULT FLUENT READERS : G RAMMATICAL PRIMING EFFECTS OF AN ROUGHLY EQUAL MAGNITUDES WERE OBSERVED IN THE LONG SOA INTERVAL CONDITION (600 MS ) AND IN THE SHORT SOA INTERVAL CONDITION (250 MS ):

23 HOWEVER, T IME WINDOWS FOR WORD PROCESSING AT 600 MS SOA AND 250 MS SOA ( ONLINE – DELAYED RT s) APPEAR TO BE SIGNIFICANTLY DIFFERENT : 125 MS VS 183 MS, RESPECTIVELY S UCH AN OUTCOME REQUIRES CAUTION IN GRAMMATICAL PRIMING EFFECTS LOCI EXPLAINING AND SPECIFYING.