Morphotactic effects on the processing of Italian derivatives Pier Marco Bertinetto Chiara Celata Luigi Talamo Scuola Normale Superiore Scuola Normale Superiore Università di Bergamo P.zza dei Cavalieri 7 P.zza dei Cavalieri 7 Piazzetta Verzieri 1 Pisa, Italy Pisa, Italy Bergamo, Italy bertinetto@sns.it celata@sns.it luigi.talamo@unibg.it morphotactic transparency. Derived forms per- Abstract taining to two different classes of morphotactic transparency but matching for length, average This paper investigates the processing of frequency, stress pattern, as well as Italian affixed forms differing for morphosemantic transparency were used as im- morphotactic transparency. A lexical de- mediate primes in a lexical decision task; the cor- cision task with immediate priming was responding underived words were used as tar- used. Following the principles of gets. Following the principles of morphotactic morphotactic transparency and Natural transparency and Natural Morphology, the prim- Morphology, the priming effect was hy- ing effect was hypothesized to be stronger for pothesized to be stronger for items with a items with a higher degree of morphotactic higher degree of morphotactic transpar- transparency. ency. However, the predictions were not totally met. The paper discusses possible 2 Morphotactic Transparency explanations from the theoretical and methodological points of view, and high- To date, the only database for morphotactic lights potential developments of the re- transparency of derivational processes is search. derIvaTario, an open-source annotated lexicon of about 11,000 Italian derivatives 1 Introduction (; see Talamo & Celata, 2011; Talamo et al., submitted). The lexical According to Dressler (1985, 2005), source of derIvaTario is CoLFIS, Corpus e morphotactic transparency is one of the main Lessico di Frequenza dell’Italiano Scritto parameters within the universal markedness the- (Bertinetto et al. 2005), a fully lemmatized three- ory of the so-called System-Independent Mor- millions word corpus of written Italian, sampled phological Naturalness. It assumes the wide- out of a carefully balanced variety of books, spread existence of “opacifying obstructions” journals and newspapers. CoLFIS was created (Dressler, 2005: 272) in inflectional, derivational with the purpose of representing the mental lexi- or compounding processes, and is expressed by con of the ideal Italian speaker – or, more exact- preference degrees along a naturalness scale. The ly, reader – as reliably as possible (Laudanna et most natural forms are those without opacifying al., 1995). obstructions, followed by those based on mildly derIvaTario takes into account several mor- opacifying phonological processes (such as phological properties of the base and of each af- resyllabification), while allomorphic rules and fix involved in the derivational cycles, crucially suppletion are the most opaque and least natural including morphotactic and morphosemantic morphological operations. In this approach, natu- transparency (see Libben, 1998 and Dressler, ral is synonymous with cognitively simple, icon- 2005 for the latter). With respect to the former, ic and therefore easy to acquire and process. derIvaTario provides a value according to the This work investigates the native speakers’ Universal Scale of Morphotactic Transparency processing of Italian affixed forms differing for (Dressler, 1985 and 2005). The scale values Copyright © by the paper’s authors. Copying permitted for private and academic purposes. In Vito Pirrelli, Claudia Marzi, Marcello Ferro (eds.): Word Structure and Word Usage. Proceedings of the NetWordS Final Conference, Pisa, March 30-April 1, 2015, published at http://ceur-ws.org 109 range from mt1 to mt8, as shown in Appendix A. The priming effect of the derivatives was as- The items used in the present experiment be- sessed as the average RT difference between the longed to two sets of derivatives, respectively morphological condition and the identity and characterized by full transparency (mt1) and rela- unrelated conditions. A statistically significant tive opacity (mt4). interaction between priming condition (morpho- logical, identity, unrelated) and morphotactic 3 Experiment transparency (mt1 vs. mt4) would suggest that the morphotactic contrast is cognitively salient. 3.1 Materials and methods Adult native Italian speakers participated in a 3.2 Results speeded lexical decision task with orthographic Repeated measure ANOVAs were run with prim- stimuli. 32 words and 32 nonwords functioned as ing condition as within-subject factor and targets. Each target (consisting of an underived morphotactic transparency as between-subject word) was immediately preceded by a prime in factor. The mean results are shown in Table 1. three different conditions: morphological (e.g. Comparing the morphological and the unrelated ribellione/ribelle, ‘rebellion/rebel’), identity conditions, mt1 primes facilitated target recogni- (ribelle/ribelle) and unrelated (xxxxxx/ribelle). tion to a larger extent than mt4 primes. Similarly, Participants saw each target in only one of the comparing the morphological condition with the three conditions. The test items are listed in Ap- identity condition, mt4 primes slowed down tar- pendix B. get recognition to a larger extent than mt1 All primes were morphosemantically fully primes. Although the general tendency was con- transparent. Half of them were classified as mt1 sistent with the experimental hypothesis, the in- according to derIvaTario (full transparency), the teraction condition x morphotactic transparency other half as mt4 (with intervening morpho- was not significant (Pillai’s trace F=0.547, p > phonological opacifying process). The two .05). Thus, although the priming effect exerted groups were carefully balanced for: (a) average by mt4 derivatives onto the corresponding un- lexical frequencies of both primes and targets, derived words was weaker than the one yielded (b) length of prime and target (as measured by N by mt1 derivatives, the current experiment does of phonemes and N of graphemes), and (c) type not support the initial hypothesis. of base. The last point needs clarification. As is well-known, Italian morphology is not word- Table 1. Average reaction times and differential prim- based, i.e. the base does not correspond to an ing (ms) across conditions and transparency levels. actual word. Since derIvaTario assumes 7 base types, it was necessary to control for the possible identity morphological unrelated effect of this variable. Only the two most fre- mt1 491 547 631 diff. 56 84 quent base types were used in the present exper- mt4 502 573 637 iment: (i) root, i.e. an underived word without diff. 71 64 inflectional ending (e.g. bellezza ‘beauty’ as based on the root bell- of bello ‘beauti- ful.M.SG.’), (ii) verbal theme, i.e. a verb root plus 4 Discussion the thematic vowel (e.g. battimento ‘beat’ as The purpose of this experiment was to investi- based on the verbal theme batti- of battere ‘to gate whether morphotactic transparency is a cog- beat’). These two base types were equally dis- nitively relevant factor in the processing of Ital- tributed within the two word sets: 11 verbal ian base forms when primed by corresponding themes, 5 roots. derivatives. A significant differential priming Nonwords were created by replacing one pho- effect was expected between mt1 and mt4 neme in real Italian derivatives and the corre- primes, which would have lent support to the sponding underived words. They had the same Universal Scale of Morphotactic Transparency as average length as the test words. implemented by derIvaTario. The experiment, The order of words and nonwords was ran- however, did not produce the expected result, domized across participants. Before performing despite encouraging tendencies. the task, the participants were trained on a list of A possible explanation for this result is the 8 items (4 words, 4 nonwords). strictly on-line character of the technique used 110 (immediate priming). As Laudanna et al. (2004) Luigi Talamo, Pier Marco Bertinetto, and Chiara have shown for verbal inflection, the effect of Celata C. (submitted) DERIVATARIO: An annotated complex morphological properties on the pro- lexicon of Italian derivatives. cessing of isolated words is more likely to be detected in off-line techniques, such as free recall Appendix A. Scale of morphotactic trans- tasks, implying a short-term and/or episodic parency. memory component. DEGREE NATURE OF PHENOMENON In addition, the assumed difference between mt1 none transparent and partially opaque derivatives in mt2 purely prosodic and phonological (e.g. resyllabification, assimilation) priming their base forms might surface to a larg- mt3 phonological, with neutralization of phonetic constitu- er extent when the morphological condition is ents (e.g. flapping) compared with a phonological priming condition mt4 morpho-phonological, without loss of constituents (e.g. articulatory weakening) (e.g. colazione/colare ‘breakfast/percolate’), in mt5 morpho-phonological, with loss of constituents which no morphological relatedness is found (e.g. deletion) between the prime and the target, although their mt6 purely morphological (e.g. paradigmatic alternation of affixes) formal relationship is the same as in a morpho- mt7 lexical: weak suppletion logically related pair (e.g. formazione/formare mt8 lexical: strong suppletion ‘formation/form’). This hypothesis is currently under investigation. Appendix B. Experimental words. Mt1 Mt4 References Prime Target Prime Target disegnatore disegnare traducibile tradurre Pier Marco Bertinetto, Cristina Burani, Alessandro bruciatore bruciare discutibile discutere Laudanna, Lucia Marconi, Daniela Ratti, Claudia suggeritore suggerire tessitura tessere Rolando, and Anna Maria Thornton. 2005. Corpus cancellazione cancellare competitore competere e Lessico di Frequenza dell’Italiano Scritto (CoL- esclamazione esclamare emettitore emettere FIS). http://linguistica.sns.it/CoLFIS/Home.htm dominazione dominare roditore rodere nuotatore nuotare scommettitore scommettere Wolfgang U. Dressler 1985. On the Predictiveness of accentuzione accentuare perseguibile perseguire Natural Morphology. Journal of Linguistics, 21(2): bollitura bollire godibile godere piegatura piegare cedimento cedere 321–337. fregatura fregare spargimento spargere Wolfgang U. Dressler 2005. Word-Formation in Nat- intrusione intruso rassegnazione rassegnato perversione perverso concitazione concitato ural Morphology. In: Štichauer P., Lieber R. (eds) ribellione ribelle desolazione desolato Handbook of Word-Formation, Springer: 267–284. introversione introverso discrezione discreto avversione avverso depravazione deparavato Alessandro Laudanna, Simone Gazzellini, and Maria de Martino 2004. Representation of grammatical properties of Italian verbs in the mental lexicon. Brain and Language 90: 95–105. Alessandro Laudanna, Anna Maria Thornton, Giorgi- na Brown, Cristina Burani, and Lucia Marconi 1995. Un corpus dell’italiano scritto contempora- neo dalla parte del ricevente. In: Bolasco S., Lebart L., Salem A. (eds) III Giornate internazionali di Analisi Statistica dei Dati Testuali, Roma: Cisu: 103–109. Gary Libben 1998. Semantic Transparency in the Processing of Compounds: Consequences for Rep- resentation, Processing, and Impairment. Brain and Language 61: 30–44. Luigi Talamo and Chiara Celata 2011. Toward a mor- phological analysis of the Italian lexicon: develop- ing tools for a corpus-based approach. Quaderni del Laboratorio di Linguistica della Scuola Nor- male Superiore di Pisa, 10. 111