Talks

Non-compositional meaning: word processing and formal linguistic theory

Paula Armelin, U. F. Juíz de Fora – Brazil (armelin.paula@ufjf.br)

Mercedes Marcilese, U. F. Juíz de Fora – Brazil (mercedes.marcilese@ufjf.br)

Cristina Name,  U. F. Juíz de Fora – Brazil & CNPq – Brazil (cristina.name@ufjf.br)

This work discusses the licensing of non-compositional interpretation triggered at the word level in a perspective that integrates psycholinguistic experimental data and formal linguistic theory, more specifically, based on the developments of the Distributed Morphology framework (Halle; Marantz, 1993, Marantz, 1997 and much subsequent work), henceforth DM. The DM model assumes a syntactic approach to word formation and is centrally based on the idea of a full decomposition of complex linguistic objects, which must be derived from the syntax. Psycholinguistic data suggest that access to individual lexical items is relevant for the computation of non-compositional meaning in idioms (Barreto; Marcilese; Oliveira, 2018; Oliveira, 2018; Barreto, 2017, for BP; Gibbs et al., 1989; Gibbs; Nayak, 1989; Gibbs; Gonzalez, 1985; Peterson et al., 1989, 2001). In this paper, we aim to investigate in which extent morphological decomposition is relevant in word comprehension when non-compositional meaning is at stake. More specifically, we adopt a cross-modal priming paradigm inspired by Cacciari and Tabossi (1988) methodology for the non-compositional investigation to seek evidence about the processing of the non-compositional meaning of words in Brazilian Portuguese (BP) formations with the diminutive suffix -inho, in data such as beijinho (with the compositional meaning of “little kiss” and non-compositional meaning of a type of candy). In an ongoing lexical decision experiment, the target word could be related

(eg. afeto “affection”), idiomatically related (eg. chocolate “chocolate”) or unrelated (eg. gravata “necktie”) to the non-compositional word. Participants listen to an utterance that provides the discursive context for the non-compositional word and immediately after that, a target word is presented on the screen for the lexical decision task. Type of context (favorable to a literal or compositional interpretation or a neuter one) and type of target word (related to literal, compositional or no related meaning) are the independent variables.

Reaction time and response accuracy are dependent variables. Priming effects are expected as a function of morphological decomposition in line with the idea that the internal structure of the word is relevant in non-compositional interpretation. From the point of view of a formal approach, the experimental data has the potential to add relevant considerations to the lexicon versus syntax debate as the word-forming component of the grammar. From a lexicalist perspective, linguistic objects, simple or complex, with non-compositional meanings are stored in an independent, pre-syntactic component – the lexicon –, which is characterized as the repository of idiosyncrasies of grammar (Chomsky, 1981; Di Sciullo & WillIams, 1987; Jackendoff, 1997, 2002, among others). However, if morphological decomposition is relevant to the interpretation of non-compositional meaning, then morphosyntactic structure should be considered in the licensing of non-compositional interpretation. This, in turn, would be compatible with the assumption of full decomposition, as the one assumed in DM, in which complex linguistic objects formed by more than one morpheme are derived by syntax, even if it involves an idiosyncratic meaning. An important issue in this perspective is to explain how syntactic structure, although compositionally constructed by the concatenation of functional heads and roots in syntax, can be non-compositionally interpreted. Our working hypothesis is that the morphosyntactic structure is present in both compositional and non-compositional interpretations. The compositional content is derived if each syntactic node receives its interpretation, considering the local structural context. The non-compositional interpretation, on the other hand, is derived if the content interpretation is attributed to a higher syntactic node (BORER, 2013, 2014) or if the syntactic nodes that interact semantically for the licensing of non-compositional content are adjacent to each other (Marantz, 2013). In Borer’s (2013, 2014) approach the non-compositional meaning constitutes the output of a unique encyclopedic search that could correspond either to simple words or to more complex structures, whereas in the contextual allosemy approach developed in Marantz (2013) the possible structurally intervening morphemes are considered to be semantically null, licensing the non-compositional interpretation.

References

• Barreto, S. O. G. (2017). Compreendemos “pintar o sete” e “pintar o quadro” da mesma forma? Um estudo experimental sobre o processamento de expressões idiomáticas no PB. Universidade Federal de Juiz de Fora: MA dissertation.

• Barreto, S. O. G., Marcilese, M., Oliveira, A. J. A. (2018). Idiomaticidade, familiaridade e informação prévia no processamento de expressões idiomáticas do PB. Letras de Hoje 53 (119-129).

• Borer, H. (2013). The syntactic domain of content. In: Becker, M., Grinstead, J., Rothman, J., eds. (2013). Generative Linguistics and Acquisition: studies in honor of Nina Hyams (205-248). Amsterdam: John Benjamins Publishing Company.

• Borer, H. (2014). Derived nominals and the domain of content. Lingua 41 (71-96).

• Cacciari, C., Tabossi, P. (1988). The comprehension of idioms. Journal of Memory and Language 27.6 (668-683).

• Chomsky, N. (1981). Lectures on Government and Binding. Dordrecht: Foris.

• Di Sciullo, A.-M., Williams, E. (1987). On the Definition of Word. Cambridge, MA: The MIT Press.

• Gibbs JR, R. W., Gonzalez, G. P. (1985). Syntactic frozenness in processing and remembering idioms. Cognition 20.3 (243-259).

• Gibbs JR, R. W., Nayak, N. P. (1989). Psycholinguistic studies on the syntactic behavior of idioms. Cognitive Psychology 21.1 (100-138).

• Gibbs JR, R. W., Nayak, N. P., Cutting, C. (1989). How to kick the bucket and not decompose: Analyzability and idiom processing. Journal of memory and language 28.5 (576-593).

• Halle, M. (1997). Distributed Morphology: impoverishment and fission. Bruening, B., Kang,

• Y., McGinnis, M., eds (1997). MIT Working Papers in Linguistics: PF: papers at the Interface 30 (425–449). Cambridge, MA: The MIT Press.

• Halle, M., Marantz, A. (1993). Distributed Morphology and the pieces of inflection. Hale, K., Keyser, S. J., eds (1993). The View From Building 20: Essays in Linguistics in Honor of Sylvain Bromberger (111-176). Cambridge, MA: The MIT Press.

• Jackendoff, R. (1997). The Architecture of the Language Faculty. Cambridge, MA: The MIT Press.

• Jackendoff, R. (2002). Foundations of Language: Brain, Meaning, Grammar and Evolution. Oxford, Nova Iorque: Oxford University Press.

• Marantz, A. (2013). Verbal argument structure: Events and participants. Lingua 130 (152-168).

• Oliveira, A. J. A. (2018). Compreensão de expressões idiomáticas do PB por falantes de línguas orais e de sinais como L1: um estudo experimental. Universidade Federal de Juiz de Fora: MA dissertation.

• Peterson, R. R., Burgess, C.; Dell, G. S.; Eberhard, K. (1989). Dissociation of syntactic and semantic analyses during idiom processing. Second Annual CUNY Conference on Human Sentence Processing, New York.

• Peterson, R. R.; Burgess, G. S. Dell; K. Eberhard. (2001). Dissociation between syntactic and semantic processing during idiom comprehension. Journal of Experimental Psychology: Learning, Memory, and Cognition 27.5 (1223-37).

The impact of morphological structure on the processing of prefix verbs

Irene Fally, University of Vienna (irene.fally@univie.ac.at)

Eva Smolka, University of Vienna (eva.smolka@univie.ac.at)

Word structure has been shown to influence lexical processing, specifically by facilitating word retrieval (Creemers et al., 2020; De Grauwe et al., 2019; Smolka et al., 2018). However, there is some evidence that sensitivity to morphological structure is strongly intertwined with other linguistic factors, like syllable structure, orthographic depth, or meaning compositionality, making it difficult to individuate specifically morphological effect (cf. Embick et al., 2021).

This is further complicated by the fact that experiments have shown variability and differences between languages, calling into question generalized models of morphological processing.

This study sets out to investigate the impact of morphological structure on lexical processing, and side-stepping confounds by comparing the processing of prefix verbs in French and Italian, two related languages that share the same word formation mechanisms (Rainer, 2008) but differ in their orthographic depth. For this study, two parallel primed lexical decision experiments with 50 L1-speakers of French and Italian respectively were conducted. In these experiments, the effect of the prior presentation of prefixed verbs (e.g. comporter ‘to behave’) to base verbs (e.g. porter ‘to carry’) on decision latencies of the base verbs is measured. The study design takes into account the semantic and morpho-orthographic relationship between the base verb and the derivate, as well as frequency, morphological family size, and orthographic neighborhood density. The experiments are set up to determine whether speakers decompose the prefixed verbs and to what extent this is facilitated by the morpho-semantic structure of the verbs. This includes the use of two measures of semantic transparency, one derived from human ratings, where L1 speakers were asked to rate word pairs on a transparency scale, and the other based on vector semantic models (Grave et al., 2018; Marelli & Baroni, 2015).

Preliminary results show that French speakers are more sensitive to morpho-semantic structure than speakers of Italian, where facilitation effects for morpho-semantically related items are generally weaker. This talk aims to explore the repercussions of these differences and implications for models of morphological processing.

References

• Creemers, A., Goodwin Davies, A., Wilder, R. J., Tamminga, M., & Embick, D. (2020). Opacity, transparency, and morphological priming: A study of prefixed verbs in Dutch. Journal of Memory and Language 110, 104055. https://doi.org/10.1016/j.jml.2019.104055

• De Grauwe, S., Lemhofer, K., & Schriefers, H. (2019). Processing derived verbs: The role of motor-relatedness and type of morphological priming. Language, Cognition and Neuroscience 34(8), 973–990. https://doi.org/10.1080/23273798.2019.1599129

• Embick, D., Creemers, A., & Goodwin Davies, A. (2021). Morphology and the mental lexicon: Three questions about decomposition. In A. Papafragou, J. C. Trueswell, & L. R. Gleitman (Eds.), The Oxford Handbook of the Mental Lexicon. Oxford University Press. 

• Grave, E., Bojanowski, P., Gupta, P., Joulin, A., & Mikolov, T. (2018, March 28). Learning Word Vectors for 157 Languages. Proceedings of the International Conference on Language Resources and Evaluation (LREC 2018). https://doi.org/10.48550/arXiv.1802.06893

• Marelli, M., & Baroni, M. (2015). Affixation in semantic space: Modeling morpheme meanings with compositional distributional semantics. Psychological Review, 122(3), 485–515. https://doi.org/10.1037/a0039267

• Rainer, F. (2008). Konvergenz- und Divergenzphänomene in der Romania: Wortbildung. In Romanische Sprachgeschichte. Ein internationales Handbuch zur Geschichte der romanischen Sprachen (pp. 3293–3307). De Gruyter.

• Smolka, E., Libben, G., & Dressler, W. U. (2018). When morphological structure overrides meaning: Evidence from German prefix and particle verbs. Language, Cognition and Neuroscience 34(5), 599–614. https://doi.org/10.1080/23273798.2018.1552006

The processing of compound nouns in bilinguals Portuguese - English: a priming study

Juliana Novo Gomes, U. Porto - Portugal

Márcio Leitão, U. F. Paraíba / LAPROL - Brazil

Marianne Duarte, U. F. Paraíba / LAPROL - Brazil

Many studies have investigated how speakers process compound nouns in their first language (Libben, 2012; Libben, Jarema, 2016; Alonso, Castellanos, Oliver, 2016). However, few studies have investigated how bilinguals process the compounds in their second language. This was the aim of the present study in which the priming technique as a methodology (Gomes & França, 2015; Garcia, 2009) was used. In it, the target was a compound noun and the prime was one of its constituents: modifier or head. We used the initial (modifier) and final (nucleus) constituents as prime and the compounds as targets. We aimed to analyze the processing of target compound nouns in three conditions: (i) transparent – tea/teacup, (ii) partially opaque – straw/ strawberry, and opaque – black/blackmail by Portuguese-English bilinguals. Besides the prime/target relationship, we tested the initial (modifier) and final (nucleus) constituents as prime and the compounds as target: (i) transparent – light/sunlight, (ii) partially opaque – fly/ butterfly, and opaque – child/brainchild. Furthermore, we aimed to determine whether there is greater facilitation in conditions in which the prime is the nucleus of the compound compared to those conditions in which the prime is the compound's modifier. We also seek to verify whether the level of proficiency will affect the performance of these bilinguals in processing the target word. We hypothesized that transparent compounds would be processed faster, followed by partially opaque and opaque compounds, however, advanced bilinguals would process compounds faster than intermediates. In addition, we expected that the head of the compound, about the modifier, would facilitate the processing of the phenomenon. Our findings were in the direction of our hypothesis. The transparent compounds were processed faster, followed by the partially opaque and opaque compounds by bilinguals. However, the groups showed different results: the advanced group showed priming effects, but, for the intermediates, there was no effect. The head, compared to the modifier, proved to be more influential in the processing of composites, at least for advanced bilinguals.

References

• Alonso, J; Castellanos, S; Oliver, M. (2016). Masked constituent priming of English compounds in native and non-native speakers. Language, Cognition and Neuroscience 39.8 (1038-1054).

• Garcia, D. (2013). Efeitos composicionais no reconhecimento visual de palavras compostas em inglês: Um estudo de MEG. PhD in Linguistics. Universidade Federal do Rio de Janeiro, Rio de Janeiro.

• Libben, G., 2012. Semantic transparency in the processing of compounds: Consequences for representation, processing, and impairment. Brain and Language 61.1 (30–44).

• Libben, G; Jarema, G. (2006). The Representation and Processing of Compound Words. Oxford University Press.

Disentangling morphological and phonological effects in the recognition of complex words in German

Anna Gupta, University of Konstanz – Germany (anna.gupta@uni-konstanz.de)

Charles Redmon, University of Oxford, University of Essex – UK (charles.redmon@ling-phil.ox.ac.uk)

Frans Plank, University of Konstanz, Somerville College (frans.plank@uni-konstanz.de)

Aditi Lahiri, University of Oxford, Somerville College (aditi.lahiri@ling-phil.ox.ac.uk)

Carsten Eulitz, University of Konstanz – Germany (carsten.eulitz@uni-konstanz.de)

Extensive empirical evidence provides support for morpheme-based theories of mental representations of complex words (Taft & Forster, 1975; Marslen-Wilson et al., 1994; etc.). One of the morphological factors that affects word processing is derivational depth defined as the number of steps used in assembling a derived word from its base. For example, the noun soaking is derived in one step from a verb (soak V → soaking N ), whereas, the noun boating, overtly containing the same number of morphemes, is a two-step derivation (boat N → boat V → boating N ). It has been demonstrated in fMRI studies on English and German (Meinzer et al., 2009; Pliatsikas et al., 2014) and behavioural studies with a focus on English conversion (Darby & Lahiri, 2016; Wheeldon et al., 2019) that there are larger processing costs for two-step words with a more complex internal structure compared to less complex one-step words.

To further explore the role of derivational depth in the recognition of morphologically complex words, we conducted two ERP studies. The goal of the first ERP experiment was twofold: first, to investigate the role of derivational depth in the processing of words with overt affixes; second, to tease apart the effects of depth and length (i.e. the number of syllables). In a cross-modal priming experiment, thirty-three German native speakers were presented with three sets of prime-target pairs in which prime depth and length were manipulated: (1) one-step disyllabic, un-weit ‘not far’ ~ WEIT ‘far’; (2) one-step trisyllabic, un-sicher ‘unsafe’ ~ SICHER ‘safe’, and (3) two-step trisyllabic, un-freund-lich ‘unfriendly’ ~ FREUND “friend”. For both experiments, we ran linear mixed-effects models on the ERP data generated by averaging voltage amplitudes across items by participant, condition, and electrode position in the time window of interest. The results of the first study demonstrated that all complex prefixed adjectives were decomposed into stems and affixes as reflected in the magnitude of priming in the N400 time window. However, the number of derivational steps and the number of syllables did not affect amplitude values. A possible explanation is that the effects of derivational depth could have been obscured by the differences in distributional properties of words under study. Despite being more complex, two-step words were morphologically more transparent, contained productive suffixes, and had higher lemma frequencies of base and whole word forms. 

In the second experiment, we aimed to disentangle the effects of derivational depth from the effects of accompanying umlaut, a morphologically conditioned phonological mutation of stem vowels. Twenty-eight native speakers of German performed a lexical decision task while their EEG activity and behavioural responses were recorded. The stimuli included three types of complex words with the nominalizing suffix –ung: (1) one-step non-umlauted, with complex nouns formed in one step from verbs (Ladung N ‘load’ ← laden V ‘to load’); (2) two-step non-umlauted, with nouns derived in two steps from nouns or adjectives via verbs (Planung N ‘planning’ ← planen V ‘to plan’ ← Plan N ‘plan’); (3) two-step umlauted, with verbs umlauted in the first step of derivation (Schwächung N ‘weakening’ ← schwächen V ‘to weaken’ ← schwach A ‘weak’). The effects of derivational depth and umlaut emerged in the N400 and LPC latency ranges. In particular, the depth effect was found for one-step non-umlauted versus two-step umlauted words in the N400 time window. The effect of umlaut appeared in the N400 range when words differed in-depth and in the LPC range when words had the same depth: voltage amplitudes were more enhanced for two-step umlauted words in both time windows which was taken as an index of increased processing demands.

Taken together, our results allow us to draw the following conclusions. First, we can confirm that word processing in German is influenced by derivational depth if one of the derivational steps involves conversion. Second, the role of depth in overt word formation requires further investigation. However, there are indications that overt derivational affixation and zero-derivation/conversion, though both contribute to complexity, do so in different ways. Third, derivational depth strongly interacts with phonological and non-morphological word characteristics, in particular, with base and whole word frequency. To sum up, a combination of morphological, phonological, and non-morphological factors affected the recognition of complex words in our studies.

References

• Darby, J. & Lahiri, A. (2016). Covert morphological structure and the processing of zero-derived words. The Mental Lexicon 11(2), 186–215. doi.org/10.1075/ml.11.2.02dar

• Marslen-Wilson, W., Tyler, L. K., Waksler, R., & Older, L. (1994). Morphology and meaning in the English mental lexicon. Psychological Review 101, 3–33. doi.org/10.1037/0033-295X.101.1.3

• Meinzer, M., Lahiri, A., Flaisch, T., Hannemann, R., & Eulitz, C. (2009). Opaque for the reader but transparent for the brain: Neural signatures of morphological complexity. Neuropsychologia 47 (8–9), 1964–1971. doi.org/10.1016/j.neuropsychologia.2009.03.008

• Pliatsikas, C., Wheeldon, L., Lahiri, A.; Hansen, P. C. (2014). Processing of zero-derived words in English: An fMRI investigation. Neuropsychologia 53, 47–53. doi.org/10.1016/j.neuropsychologia.2013.11.003

• Taft, M.; Forster, K. I. (1975). Lexical storage and retrieval of prefixed words. Journal of Verbal Learning and Verbal Behavior 14(6), 638–647. doi.org/10.1016/S0022-5371(75)80051-X

• Wheeldon, L., Schuster, S., Pliatsikas, C., Malpass, D., & Lahiri, A. (2019). Beyond decomposition: Processing zero-derivations in English visual word recognition. Cortex 116, 176–191. /doi.org/10.1016/j.cortex.2018.09.003

​

An experimental analysis of idiomaticity in prefixed words in Brazilian Portuguese 

Isabella Lopes Pederneira, U. F. Rio de Janeiro – Brazil (isabellapederneira@letras.ufrj.br)

Thays Ferreira Alves, U. F. Rio de Janeiro – Brazil (thays.ferreira37@letras.ufrj.br)

The objective of this work is to relate the theory of grammar with the explanation for idiomatization in Brazilian Portuguese (BP) words, based on experimental results. The assumed grammar model is the Exoskeletal model (BORER, 2003), according to which syntactic computation goes to the interior of words. In this theoretical framework, idiomaticity in words can occur at any layer during their derivation, explaining results from experiments carried out with multilayered prefixed words in BP. Pederneira (2010) observed that a prefix can idiomatize words, which causes a space between the structural derivation and the phonological demarcation. Experimental results showed that prefixes in words in which there is a gap in their compositional meaning can cause changes in the perception of the morphological family and possibly give rise to new roots, e.g., despencar (fall), in which we no longer compose with the name penca (bunch). With the loss of the syntactic composition of the base word, the derivatives with one more layer become the first, and a new root is created, with the first syllable just similar to the prefix. We can also ask whether there is a limit to prefix idiomaticity and confirm theoretical predictions from another constructionist model, Distributed Morphology (Marantz, 1997). We consider Marantz (2007), who proposes a limit to the idiomaticity for the prefix re-, and Medeiros (2016) who, extending the debate to the prefix -des in BP, addressed the recursiveness of these prefixes as in re-remeter (re-re-put) and des-desmarcar (un-un-cancel). We will show experimental results for cases with more than one prefixal layer, such as comprometer (commit), which, unlike the prefix re-, idiomatizes even after the insertion of a second prefixal layer (com+pro). The hypothesis is that equal prefixes are like recursion; conversely, different prefixes free the compound for new, potentially infinite idiomatizations. We took advantage of the improvement of psycholinguistic techniques (Maia; Lemle; França, 2007) about morphological processing in Portuguese. Therefore, we used the psycholinguistic Lexical Access protocol with covert priming, in which priming stimulation was performed with prime/target, non-prime/target pairs, and also prime/non-word pairs. Primes first appear on the screen and then the targets appear. Covert priming is on the screen for only 38ms, to stay below the threshold of consciousness, the participant's task is to decide whether the target item is a word or not.

The conditions are words with only one prefix, words with two equal prefixes, and words with two different prefixes, in addition, for each condition there is compositional and idiomatic reading, as well as control conditions. As independent variables, we have the morphological and semantic relationship, and as dependent variables, we have the decision indices and response time for the “yes” options in the priming test with the lexical decision. By following the experimental literature of Longtin (2003) and Marslen Wilson (1994), it is possible to observe the semantic consequence of the internal structure of words with prefixes. The experimental results show less speculative tools for the theoretical explanation of such diverse meanings for the same root with multiple prefix layers.

References

• Borer, H. (2003). Exo-skeletal vs. Endo-skeletal explanations: syntactic projections and the lexicon. In: Moore, John; Polinsky, Maria (ed.). The Nature of Explanation in Linguistic Theory (31–67). Chicago: University of Chicago Press (CSLI).

• Longtin, C. M.; Segui, J.; & Hall, P. A. (2003). Morphological priming without morphological relationship. Language and Cognitive Processes 18.3, (313-334).

• Maia, M.; Lemle, M.; França, A. I. (2007). Efeito stroop e rastreamento ocular no processamento de palavras. Ciência & cognição 12 (2-17).

• Marantz, A. (1997). No escape from syntax: Don’t try a morphological analysis in the privacy of your own lexicon. In: Dimitriadis, A.; Siegel, L. et al., eds. University of Pennsylvania Working Papers in Linguistics 4.2 (201-225).

• Marantz, A. (2007). Restitutive re- and the first phase syntax/semantics of the VP. New York: New York University. Unpublished manuscript.

• Marslen-Wilson, W.; Tyler, L. K.; Waksler, R., Older, L. (1994). Morphology and meaning in the English mental lexicon. Psychological Review 101.1 ( 3-33).

• Medeiros, A. B. (2016). Prefixos, recursividade e a estrutura do sintagma verbal. Revista do GEL 13.1 (56–86). DOI: 10.21165/gel.v13i1.541.

• Pederneira, I, L. (2010). Etimologia e reanálise de palavras. MA Dissertation. U. F. R. J.

MorphoPlay: assessing word morphological knowledge in European Portuguese children to identify and support literacy learners[1]

Carina Pinto, IP Leiria – Portugal (pintocarinaal@gmail.com)

Alina Villalva, U. Lisbon / The Word Lab – Portugal (alinavillalva@edu.ulisboa.pt)

Rafael Dias Minussi, U. F. São Paulo / LabLinC / The Word Lab – Brazil (rafael.minussi@unifesp.br)

Cândida Silva, IP Leiria - Portugal

Etelvina Lima, IP Leiria - Portugal

Morphological knowledge concerns the capacity to understand the structure of words, which requires the ability to recognize all morphological constituents and understand their interrelations. This understanding is crucial for forming meaningful words that can be appropriately used in sentences (Amenta & Crepaldi, 2012; Goodwin et al., 2020). Several studies (cf. Apel et al., n.d.; Carlisle, 2000; Varga et al., 2022) demonstrate that morphological knowledge is a vital component of literacy, playing a pivotal role in children’s successful development of reading and writing skills, beneficiating children’s linguistic proficiency in several aspects:

1. Word Decoding: Morphological awareness provides children with valuable cues to understand words. For instance, recognizing that the suffix -eiro often denotes profession names allows a child to decode the term mineiro ‘miner’; as ‘a person who works in a mine’. Further recognizing that the same suffix denotes plant names when the base is a fruit name allows the child to decode pessegueiro ‘peach tree’ as a ‘tree that produces peaches’.

2. Fluent Reading: Expediting word recognition, morphological knowledge contributes to enhanced reading fluency. For instance, familiarity with the relationship between celebração ‘celebration’; and celebrar ‘celebrate’; enables swift recognition of the former.

Despite its significance, numerous children face challenges with morphological features, hindering their ability to leverage this knowledge in word production, understanding, and reading comprehension. Consequently, the assessment of morphological knowledge becomes essential in identifying children at risk of reading and writing difficulties, facilitating the implementation of appropriate interventions (cf. Marks et al., 2022; Melloni & Vender, 2022).

The assessment of children’s morphological knowledge involves various methodologies, with morphological awareness tasks being a prevalent approach. These tasks entail manipulating morphological constituents to form new words or alter the meaning of existing ones. For example, a child may be assigned to add the suffix -ção to the verb celebrar ‘celebrate’, resulting in the complex word celebração ‘celebration’ (cf. Goodwin et al. 2020a; Goodwin, 2020b; Goodwin, 2021; Varga et al., 2022). Nevertheless, other tasks are available, and some may provide better qualitative results or more righteous quantitative data.

MorphoPlay is an ongoing project designed to explore a digital serious game as an assessment tool. In this interactive experience, players, specifically children aged 7-10, engage in puzzle-solving activities while their morphological knowledge of words is evaluated. Within the game, the child plays a small dog helping his grandfather in town activities.

The development of this project is progressing through several steps:

1. Selection of stimuli

A set of 100 words has been initially defined, comprising 20 words for each of the following conditions:

Agentive nouns in -eir(o/a) e.g., sapateiro ‘shoemaker’

Plant nouns with -eir(o/a) e.g., tomateiro ‘tomato tree’

Deverbal agentive nouns in -dor(/a) e.g., patinador ‘skater’

Unattested deverbal agentive nouns in -dor(/a) e.g., construidor ‘build + dor’

Simplex agentive nouns e.g., advogado ‘lawyer’

2. Word Association Task

The preceding stimuli are subjected to a WAT with a sample of 252 children. Results are categorized using a grid, differentiating morphosemantically related responses, other types of semantically related responses, otherwise related responses, and non-validated answers. Preliminary analysis reveals that agentive nouns in -eir(o/a), deverbal agentive nouns in -dor(/a) both attested and not attested are more effectively elicit words with a morphosemantic relationship than the other conditions.

3. Lexical Decision Task

The results collected from a sample of 287 children show that with literacy progresses, mean reaction times consistently decrease, showing significant differences between grades (Z=2823, 623; p< 0,001), notably with the 2 nd grade exhibiting double or more the reaction time of later years. Agent nouns derived from compositions are consistently processed faster across all grades, supporting previous research on derivative suffixes, while -eiro derivatives suggest polysemy slows processing, particularly with plant name nouns.

4. Serious game development (partially concluded)

The software and graphics for the game are now concluded. The game features four linguistic tasks utilizing a subset of the previously tested stimuli (WAT and LDT):

- a lexical decision task with reaction time measurement;

- a task where children must find the base word hidden within complex words;

- a task requiring the identification of morphologically related words;

- a task where children must identify words that do not belong to the set and are not morphologically related.

In conclusion, the assessment of word morphological knowledge in children serves as a crucial step in identifying those at risk of reading and writing difficulties and facilitating targeted interventions. MorphoPlay emerges as a promising and dependable tool poised to fulfill this purpose, catering to both speech and language therapists and language teachers.

References

• Amenta, S., & Crepaldi, D. (2012). Morphological processing as we know it: an analytical review of morphological effects in visual word identification. Frontiers in Psychology 3. 232. https://doi.org/10.3389/fpsyg.2012.00232.

• Apel, K., Petscher, Y., &Henbest, V. S. (undated). MATRS Technical Manual 1 Morphological Awareness Test for Reading and Spelling (MATRS). Technical Report. https://doi.org/10.31234/osf.io/ty2pe.

• Carlisle, J. F. (2000). Awareness of the structure and meaning of morphologically complex words : Impact on reading. Reading and Writing 12 (169–190). https://doi.org/10.1023/A:1008131926604.

• Goodwin, A. P., Petscher, Y., Jones, S., McFadden, S., Reynolds, D., & Lantos, T. (2020). The Monster in the Classroom: Assessing Language to Inform Instruction. Reading Teacher 73.5 (603–616). https://doi.org/10.1002/trtr.1870.

• Goodwin, A., Petscher, Y., & Tock, J. (2021). Multidimensional morphological assessment for middle school students. Journal of Research in Reading 44.1 (70–89). https://doi.org/10.1111/1467-9817.12335.

• Marks, R. A., Eggleston, R. L., Sun, X., Yu, C. L., Zhang, K., Nickerson, N., Hu, X. S., & Kovelman, I. (2022). The neurocognitive basis of morphological processing in typical and impaired readers. Annals of Dyslexia 72.2 (361–383). https://doi.org/10.1007/s11881-021-00239-9.

• Melloni, C., & Vender, M. (2022). Morphological awareness in developmental dyslexia: Playing with nonwords in a morphologically rich language. PLoS ONE 17.11 doi.org/10.1371/journal.pone.0276643.

• Varga, S., Pásztor, A., & Stekács, J. (2022). Online Assessment of Morphological Awareness in Grades 2–4: Its Development and Relation to Reading Comprehension. Journal of Intelligence 10.3. (47) doi.org/10.3390/jintelligence10030047.

​

Morphological processing of ambiguous compounds: Evidence from Swedish

Linda Sandström, Umeå University – Sweden (linda.sandstrom@umu.se)

Maria Rosenberg, Umeå University – Sweden (maria.rosenberg@umu.se)

Compounding is a productive process in Swedish as well as in many other languages, meaning that we likely encounter new compound word formations on a regular basis. For a majority of existing compounds, there is only one morphological structure (i.e. one decomposition possibility); however, in some cases, the morpheme boundary is blurred. Since Swedish compounds are always written as single unspaced strings, there is no way of knowing from orthography alone whether a compound like glasskål should be parsed as glass + skål (‘ice-cream’ + ‘bowl’), glas + skål (‘glass’ + ‘bowl’), or perhaps even as glass +kål (‘ice-cream’ + ‘cabbage’). Libben (2020) refers to such compounds as being in a lexical superstate, a hierarchical morphological parsing state that is thought to accommodate the whole word, as well as all possible morphological structures.

Most likely, the chosen analysis of a compound like glasskål is guided by semantics and contextual factors. But, the question remains: how do we process ambiguous compounds encountered for the very first time? When we lack existing whole-word representations and context to guide our initial parsing, do we automatically activate all possible structures, as indicated by the findings in Libben (2020)? Or is the initial parsing influenced by factors such as embedded word length, embedded word frequency, and embedded word frequency as a constituent in a compound, as suggested by Beyersmann and Grainger (2023)?

The present study aims to explore these questions. More precisely, using online masked LDT experiments (Angele et al., 2023), we investigate how readers of Swedish process novel ambiguous compounds during the early automatic stages of reading. If indeed all conceivable morphological states are activated in early processing, novel ambiguous compounds should prime all possible embedded words. On the other hand, if factors such as embedded word length and frequency affect early processing, we would expect to see a preferred morphological decomposition based on these factors. The experimental primes comprise 24 novel Swedish compounds with ambiguous morpheme boundaries (cf. Libben’s 2020 clampeel with possible parsing into clamp + eel or clam + peel). Four separate word lists were created such that each prime would appear once in a prime-target pair with one of the four embedded words (prime-condition 1). The four lists were matched on target length and target position within the compound. Another 24 novel unambiguous compounds were used as unrelated primes (prime-condition 2). Additionally, 24 novel unambiguous compounds with related or unrelated target words were used as control stimuli (prime-condition 3) together with 12 nonword+word combinations (prime-condition 4) and 12 word+nonword combinations (prime-condition 5). For the LDT, 96 corresponding nonword targets with related noncompound primes were created. The nonword prime-target pairs were similarly distributed across the four lists as in the word condition. See Table 1 for an example of the experiment list setup.

Table 1

An example of prime-target pairs in the four experimental lists, and an overview of the total number of prime–target pairs in the word condition.

Note. Prime condition: 1 = related ambiguous novel compounds, 2 = unrelated unambiguous novel compounds, 3 = related/unrelated unambiguous novel compounds, 4 = related/unrelated noncompounds (nonword + word), 5 = related/unrelated noncompounds (word + nonword). 

Response times and accuracy rates are analyzed as dependent measures using linear mixed-effects models. Variables considered in the analyses are: prime condition (ambiguous vs. unambiguous), prime type (related vs. unrelated), target frequency and length, and age of acquisition estimates (Brysbaert & Biemiller, 2017). To further explore more in detail the processing of the experimental ambiguous primes, the following factors are considered: target frequency as a constituent in existing compounds, embedded word length (long vs. short), and preferred analysis of ambiguity (factorial 0-1, based on ratings from 56 participants).

In conclusion, this study intends to shed light on how readers process morphologically complex words and how non-morphological features might intervene in this processing.

References

• Angele, B., Baciero, A., Gómez, P. & Perea, M. (2023). Does online masked priming pass the test? The effects of prime exposure duration on masked identity priming. Behavior Research Methods 55.1 (151–167).

• Beyersmann, E. & Grainger, J. (2023). The role of embedded words and morphemes in reading. In D. Crepaldi, ed. (2023). Linguistic Morphology in the Mind and Brain (26–49). London, Routledge.

• Brysbaert, M. & Biemiller, A. (2017). Test-based age-of-acquisition norms for 44 thousand English word meanings. Behavior research methods 49 (1520–1523).

• Libben, G. (2020). What can we learn from novel compounds? In S. Schulte im Walde & E. Smolka (eds.), The role of constituents in multiword expressions: An interdisciplinary, cross-lingual perspective (107–127). Language Science Press.

Theoretical hypothesis under the scrutiny of experimental results: the morphological nature of blends

Alina Villalva, U. Lisbon / The Word Lab – Portugal (alinavillalva@edu.ulisboa.pt)

Rafael Dias Minussi, U. F. São Paulo/LabLinC/The Word Lab – Brazil (rafael.minussi@unifesp.br)

José Ferrari Neto, U. F. Paraíba / LAPROL – Brazil (joseferrarin1974@gmail.com)

Gustavo Estivalet, U. F. Paraíba / LAPROL – Brazil (gustavoestivalet@hotmail.com)

The B&C-ProAlex ongoing research project delves into the analysis and processing of Portuguese blends and compound words. Blends present intriguing linguistic challenges for various reasons, since they defy conventional morphological analysis, sparking discussions about their intricate relationship with derived and compound words (Kubozono, 1989; López Rúa, 2002; Renner, 2022). They also test the state-of-the-art in lexical processing and access, due to their exceptionally low frequency, offering a unique opportunity to scrutinize non-lexicalized words.

The construction of a Portuguese blend corpus, comprising 423 forms, has spurred meticulous analysis (cf. Villalva & Minussi, 2022). Blends were classified according to their frequency and the frequency of their base words (Corpus do Português), and the word class they and their word constituents belong to. Furthermore, blends were classified according to:

(i) their constituent structure - blend constituents can be clips (C), which are randomly truncated forms, or words (W), which may manifest as roots, stems, or fully inflected forms; therefore, four different blend constituent structures are available - CC, CW, WC, WW;

(ii) their grammatical relationship structure - blend constituents may be heads (H) or modifiers (M); therefore, they may have a coordination structure (HH), or a head-initial (HM) or head-final (MH) modification structure;

(iii) their prosodic structure - considering the number of syllables, blends develop a prosodic relationship with the 1 st base (1b), the 2 nd (2b), both (12b), or none (øb).

The analysis of this corpus revealed that not all structural combinations are equally represented, and it unveiled interesting constraints. Almost half of the blends display a coincidence between a head constituent that is a prosodically preponderant clip. We have termed these as congruent blends (cf. cartomente, manifestoche, tristemunho, apertamento, suchá, namorido, chafé, gachorro). Disregarding the prosodic factor, since many blends show no prosodic relation to either of their base words, congruent blends increase to 70%. Less than 10% of the blends, that we have termed incongruent blends, display either a prosodically predominant clipped modifier or a mismatch between prosodic predominance and clipping (cf. cowboio, aristogato, vignorar). This set can also be expanded if the prosodic predominance is disregarded (cf. mariavilhosa, inutensílio). The remaining 20% of the blends in the corpus share a peculiar feature - both base words are visible because their internal boundaries overlap (cf. sonolentidão = sonolent + lentidão). Some of these WW blends may have a deep CW or a WC constituent structure, but it cannot be independently established (e.g., librianjo, familícia, maltrimónio, sonolentidão). The remaining WW structures are subdivided into 2 groups. The first includes relabeled incongruent structures (cf. ininterruptor), and the second is a strictly neutral structure (cf. nuventania - HH/WW/øb). The compound corpus includes more than 300 items, including morphological and morphosyntactic compounds. The first split into coordinated (luso-francês) and head-final modification structures (bafômetro). The second includes coordinated (bar-restaurante) and modification head-initial structures (bomba-relógio), as well as a particular VN structure (beija-flor). Their frequency and presence in dictionaries have also been registered.

To assess the robustness of this theoretical analysis, we conducted offline and online experiments. The offline experiment was a word association test. It was conducted through 4 Google Forms, each containing about 50 words divided between blends and compounds. Participants were asked to write down only one word, the first that came to mind. The results of this test indicate that compounds and blends are interpreted differently by the participants. While for blends, participants tend to mention one of the bases, for compounds, the highest percentage of responses did not consider the compound bases but rather another word with or without semantic relation, suggesting that blends are interpreted compositionally, and compounds are interpreted non-compositionally. Online experiments were lexical decision tests. The first compared blends to derived and simplex words, and pseudowords. Accuracy and reaction time results suggest that blends are processed differently from derived words, and similarly to pseudowords. The second experiment compared blends to compounds, simplex words and pseudowords. Accuracy and reaction time results followed the pattern of the previous test, suggesting that the processing of blends resembles the processing of pseudowords, and differs significantly from the processing of simplex and compound words, independently of their morphological structure. Furthermore, MRT for compounds is significantly larger than for simplex words, suggesting that compound processing requires a more complex decomposition and lexical access.

In conclusion, our research sheds light on the morphological nature of blends, providing a thorough analysis of their structural variations and offering experimental evidence supporting our theoretical hypotheses. The findings contribute to the broader understanding of lexical processing and access, particularly in the context of non-lexicalized words like blends.

References

• Kubozono, H. (1989). Syntactic and rhythmic effects on downstep in Japanese. Phonology 6 (39–67).

• López Rúa, P. (2002). On the structure of acronyms and neighbouring categories: a prototype-based account. English Language and Linguistics 6.1 (31–60). doi: 10.1017/S136067430200103X

• Renner V. (2022). Blending, in P. Ackema, S. Bendjaballah, E. Bonet & A. Fábregas (eds.), Wiley Blackwell Companion to Morphology. Hoboken, NJ: Wiley.

• Villalva. A; Minussi, R. D (2022). Description and analysis of a Portuguese blend corpus. Corpus 23. doi: doi.org/10.4000/corpus.6436.

[1] This work was funded by Portuguese national funds provided by Fundação para a Ciência e Tecnologia - FCT (2022.02585.PTDC) - https://doi.org/10.54499/2022.02585.PTDC.

Posters

The time course of word recognition in Brazilian Portuguese: a priming lexical study comparing skilled adult readers and developing 4th-grade readers

Daniela Cid Garcia, U. F. Rio de Janeiro - Brazil

Juliana Novo Gomes, U. Porto - Portugal

Katia Abreu, U. E. Rio de Janeiro - Brazil

The morphological structure of complex words can impact on how they are processed during visual word recognition. A word may be recognized faster if a morphologically related word is recently processed. For instance, the recognition of darkness in a priming paradigm (in which subjects read target words after prime words) is speeded by prior presentation of the prime dark. Nevertheless, priming studies have shown different patterns of facilitation revealing several lexical properties on the recognition of words that influence processing (for a review, see Amenta & Crepaldi, 2012; Diependaele, Grainger, & Sandra, 2012; Hay & Baayen, 2005; Marelli, Traficante, & Burani, 2020; Milin, Smolka, & Feldman, 2018). In the last few decades, a large number of studies has been investigating whether the facilitation can be attributed to morphology alone or if it is an effect of form and meaning overlaping. The main goal of this ongoing study is to use a priming paradigm to investigate the activation of morphologically complex words in Brazilian Portuguese (BP) as a means to verify whether the activation of a word requires its decomposition into morphemes. Also, we aim to compare skilled adult readers to developing 4 th -grade readers in order to analyze the role of morphological mapping onto visual word recognition in the first years of learning to read. According to Perfetti (2007), exposure to written material results in better lexical quality, which is the quality of the representation of a lexical item in the mental lexicon. It involves, for example, phonological, orthographic, morphological, and semantic information. Skilled readers are those who have automated decoding between graphemes and sounds, making a direct mapping between spelling and meaning. The decoding process is essential for learning to read, and even skilled readers still use phonological mapping (Ziegler; Jacobs; Kluppel, 2001). On the other hand, there is also evidence that, as readers progress to become more skilled, there is less alphabetical decoding and readers begin to automatically recognize the graphic form of words (CASTLES; RASTLE, 2018). Morphological relationships may be an important part in acquiring this direct mapping between written words and their meanings (Rastle, 2018). Garcia, Maia and França (2012) reported a masked priming lexical decision experiment, in which the target was kept constant and the prime varied so as to meet three different relations to it: (1) morphologically related -FILA/fileira (line/row); (2) semantically related only - ORDEM/fileira (order/row); (3) phonologically related only: FILÉ/fileira (steak/row) and (4) Unrelated (baseline): MATO/fileira (weed/row). The semantically-related pairs had no morphological relationship whatsoever and the approach they used to measure the level of semantic similarity followed Gomes (2009) and Gomes & França (2015) — they selected prime-target pairs that were very close semantically in the way that there was always the possibility of creating an abstract minimal syntactic structure between them, such as the pair ORDEM (order) and fileira (row) that can be instantly inserted in a prepositional phrase (i.e. ordem EM fileira: order of the row). The authors reported response times (RTs) and accuracy rates. The results were compared to those of the baseline condition and morphologically related pairs obtained the shortest response times. Here, we decided to replicate the masked priming lexical decision experiment from Garcia, Maia and França (2012) in two groups of participants: skilled adult readers and developing 4 th -grade readers. First, we validated our stimuli by asking 50 subjects to rate on a 7-point scale the degree to which prime and target were related in meaning (1=very similar; 7=unrelated). Then, the experimental stimuli were created in four lists with 16 primetarget pairs controlled for length and frequency. We assessed the priming effects upon words following the four conditions aforementioned. Each trial begins with a fixation cross on the screen for 1500ms, immediately followed by a pre-prime consisting of a sequence of stars (******), which remained on the screen for 500ms before the prime was presented on the screen (in capital letters). Prime remained for 38ms and was followed by another sequence of stars (******), which remained on the screen for 50ms and was followed by the target (lower case). Target remained on the screen for 200ms and, after that, the subject had 1500ms to perform the lexical decision. First results suggest that morphological information is faster processed and phonological information may indeed be represented or retrieved dissociated from morphological information. We hope to demonstrate the importance of morphology when it comes to the derivation of multimorphemic words, which has proved not to occur through the mere association of semantic and phonological similarity.

References

• Garcia, D. C.; Maia, M. A. R.; França, A. I. The time course of word recognition: evidence from Brazilian Portuguese. ReVEL, v. 10, n. 18, 2012. [www.revel.inf.br].

• Gomes, J. N. A Direcionalidade no Relacionamento Semântico: um Estudo de ERP. Dissertação de Mestrado em Linguística, UFRJ, 2005.

• Rastle. K.; Davis, M.D.; Marslen-Wilson, W.D.; Tyler, L.K. Morphological and semantic effects in visual word recognition: A time-course study. Language and Cognitive Processes, v. 15, 2000. 17.

• Rastle, K.; Davis, M.D.; New, B. The broth in my brother’s brothel: Morphoorthographic segmentation in visual word recognition. Psychonomic Bulletin and Review, v. 11, 2004.

The Root Meaning of Polysemous Verbs in Brazilian Portuguese: a priming self-paced reading study

Juliana Novo Gomes, U. Porto - Portugal

Isabella Lopes Pederneira, U. F. Rio de Janeiro - Brazil

The analysis of polysemous verbs can contribute to improving approaches that coexist in Generativism. For instance, Lemle and Pederneira (2011) put forward an analysis comparing argument structure of cognate verbs in BP and in Italian. Pederneira (2017) supports that classes and meanings do not stem from the potentiality of the lexical content intrinsic to a root/word or to a verb, but from a syntactic context, following predictions of the MD and the Exoskeletal model. Observing verbs in BP, such as pegar (to take) or matar (to kill), it is noticeable the variety of meanings related to different syntactic contexts – pegar o cobertor (to take the blanket), pegar uma gripe (to catch a flu), pegar aquele rapaz (to get the guy), pegar a rua Castro Alves (to go on Castro Alves street); matar passarinho (to kill a bird), matar a saudade (to appease the longing), matar a garrafa de cerveja (to chug a bottle of beer), matar a bola no peito (to chest a soccer ball). The main goal of this ongoing study is to ran a priming task combined with a self-paced reading (SPR) a behavioral tool to test two competing theoretical approaches to the relationship between Syntax and Lexicon — (i) the Lexicalist Hypothesis of Generative Grammar (Chomsky, 1970) and (ii) the Constructionist/Constructivist Hypothesis of Generative Grammar (for instance, the Distributed Morphology (MD, Marantz, 1997) and the Exoskeletal model (Borer, 2003, 2013)). These competing theories make different claims about the relationship between Syntax and Lexicon and, therefore make different predictions about sentence processing under priming relations: (i) assumes that the Lexicon corresponds to an independent module that provides the raw material for the syntactic processor. In this strict configurational model of argument structure, compositional semantics is directly read off the syntactic structure. While in (ii), the syntactic atoms corresponding to the root nodes are associated with open class exponents without a specific meaning that can select a syntactic licensing context. The lexical meaning arises constructionally, an analysis in tune with the template proposed by the Exoskeletal model.

An interesting feature about verbs (Grimshaw and Mester, 1988; Butt, 2010) is that for intransitive verbs, the locality threshold for meaning negotiation would be word-level verbal categorization [Root + v]v; while transitive verbs would have the locality limit for the meaning negotiation inside the verbal complex including the complement [[Root +v] + DP]vP. Another interesting feature is that light verbs need to be in an event structure and, from there, the emergence of meaning is possible. Hence, polysemous verbs would behave in the syntax-semantic interface with similar characteristics to light verbs: lacking semantic properties and needing a complement to form the vP giving rise to meaning.

We use word-priming paradigm followed by self-paced-reading sentences and language comprehension question study over PCibex. We tested whether the polysemous verbs inserted in different syntactic-semantic structures would be primed by related (shared roots) and unrelated words. Aiming to understand the conceptual content of the roots and their relation to the syntactic structure, we designed four conditions: (a.) related-prime w. transitive +animate structure; (b.) unrelated-prime w. transitive +animate structure; (c.) related-prime w. transitive -animate structure and, (d.) unrelated-prime w. transitive -animate.

The methods used here assume that ‘the longer and more complex the linguistic computations necessary to generate the representation — the longer it should take for a subject to perform any task involving the representation’ (Marantz, 2005, p. 439).

The novelty of our SPR-priming study is that it allows us to measure reading times and to predict faster readings for target sentences if there is lexical overlap between prime and main- verbal-head (sentence), such as in (a) and (c); while morphologically unrelated primes, such as in (b) and (d), would not show prime effects. Since the two models predict different priming effects between conditions (a) and (c), as well as, in their interaction with the rest of conditions, we aim to find evidence to restrict the basic semantic differences in the polysemous verbs in Brazilian Portuguese to their syntactic configurations that can potentially give rise to meanings.

References

• Borer, H. (2003).Exo-skeletal vs. endo-skeletal explanations: Syntactic projections and the lexicon. The nature of explanation in linguistic theory, v. 31, p. 67

• Borer, H. (2013). Structuring sense. Volume III: taking form. Oxford University Press.

• Butt, M. (2010). The light verb jungle: still hacking away. In: Amberber, M., Grimshaw, J.; Mester, A. (1988). Light verbs and theta-marking. Linguistic Inquiry, Massachusetts, v. 19, p. 205-232.

• Lemle, M; Pederneira, I.L (2011). Learning about the syntax-semantics interface: a study of cognate verbs in Brazilian Portuguese and Italian. Revista da ABRALIN.

• Maranz, A. (1997). No escape from syntax: don’t try morphological analysis in the privacy of your own lexicon, in A. Dimitriadis, L. Siegel et al., eds. University of Pennsylvania Working Papers in Linguistics, vol. 4.2, Proceedings of the 21st Annual Penn Linguistics Colloquium (201-225).

• Pederneira, I. L. (2017). Implicações Teóricas dos Verbos Leves para o Estudo de Estrutura Argumental. Veredas 20 (160-182).

Lexical access and visual word processing mediated by morphological constituents in children with language disorders[1]

Carina Pinto, IP Leiria - Portugal

Catarina Lavos, IP Leiria - Portugal

Marta Sousa, IP Leiria - Portugal

Raquel Espírito Santo, IP Leiria – Portugal

Lexical access is the process of retrieving a word from the mental lexicon. It is a complex process involving multiple representation levels, including phonological, semantic, morphological, syntactic, and orthographic (Amenta et al., 2011; Taft, 1988). Children with language disorders often have difficulty accessing words from the lexicon. This difficulty can be seen in several ways, including slow naming speed, difficulty retrieving words from long-term memory, and difficulty understanding the meaning of words (Rothou & Padeliadu, 2019). There are numerous possible explanations for this difficulty, including reduced lexical Knowledge, impaired phonological processing, and impaired semantic processing, among others (Bitan et. al., 2020). Visual word processing mediated by morphological constituents is the process of using morphemes to help access and understand words (Bitan et. al., 2020, Giazitzidou & Padeliadu, 2022). For example, knowing the meaning of the morpheme -ed can help a child understand the past tense of verbs in English. In this work, we intend to demonstrate that, in European Portuguese, children with oral and written language impairment rely on morphological constituents to access the lexicon. In a sample of 30 children, from 2 nd to 4 th grade we tested, via lexical decision task, a set of 100 words. At the moment we have a total of 15 children, out of the 30 we intend to collect (m age =8.4 ±1.121) Seven are male and wight female distributed by school year as follows: 4 in the 2nd grade, 4 th in the 3 rd , and 7 in the 4 th grade. We have different language pathologies; Language Impairment, Dyslexia, and Dysortography).

The words used are distributed by the following conditions:

1. Compositional denominal nouns in -eir(o/a) (e.g., sapateiro ‘shoemaker’)

2. Compositional denominal plant names in -eir(o/a) (e.g., tomateiro / laranjeira; ‘tomato tree’ / ‘orange tree’)

3. Compositional derivative nouns in -dor(ø/a) (e.g., patinador ‘skater’)

4. Unattested deverbal nouns in -dor(ø/a) (e.g., chorador / ‘crier’)

5. Simplex agentive nouns (e.g., acrobata ‘acrobat’)

6. Pseudowords (e.g. asepa)

The preliminary results show significant differences between the 6 conditions (Z=12,39; p<0,001). The main difference is between compositional derivatives deverbals nouns in -dor(ø/a) (e.g., patinador; skater) that contrast with all the other conditions. Compositional denominal nouns in -eir(o/a) (e.g. sapateiro ‘shoemaker’) also contrasts with almost all the other conditions, except compositional denominal plan names in –eir(o/a). Agentive nouns with a simple morphological structure and unattested deverbal nouns in -dor(ø/a), not attested also contrast with pseudowords. The results from plant names in –eir(o/a) (e.g., tomateiro ‘tomato tree’ / laranjeira ‘orange tree’) are not clear at this moment and may be due to the sample distribution or inherent properties of these words.

The preliminary results also show that frequency has a role in the visual word processing of words, even in children with language pathologies (t=4,128; p<0,001): Lexical access and morphological knowledge are important skills for language comprehension and production. Research on these skills in children with language disorders can help us to develop better interventions to improve their language abilities.

References

• Bitan, T., Weiss, Y., Katzir, T. & Truzman, T. (2020). Morphological decomposition compensates for imperfections in phonological decoding. Neural evidence from typical and dyslexic readers of a opaque orthography. Cortex 130 (172-191).

• Giazitzidou, S., & Padeliadu, S. (2022). Contribution of morphological awareness to reading fluency of children with and without dyslexia: evidence from a transparent orthography. Annals of Dyslexia 72 (509-531).

• Rothou K.M. & Padeliadu S. (2019) Morphological processing influences on dyslexia in Greek-speaking children. Annals of Dyslexia 69 (261-278).

[1] This work was funded by Portuguese national funds provided by Fundação para a Ciência e Tecnologia – FCT (2022.02585.PTDC) https://doi.org/10.54499/2022.02585.PTDC