Producing regularly and irregularly inflected verb forms: behavioural and neuroimaging data from the three Italian conjugations

ABSTRACT

The generation of regular and irregular inflected verb-forms has been taken as a crucial test between models of inflection in the mental lexicon, namely approaches which invoke different mechanisms for regular and irregular forms (rule-based procedures vs. recovery of whole forms stored in an associative network) and accounts which postulate a single procedure for all forms. An alternative hypothesis suggests that inflectional processes are explained by membership of words to clusters, for instance, Italian inflectional classes (conjugations). In a behavioural and a rapid event-related fMRI experiment, participants overtly generated the past participle of verbs from the three Italian conjugations. Results showed that the cognitive operations and the neural substrate underlying inflectional processes rely on specific properties of inflectional classes. Different patterns of cortical activations elicited by verbs from different conjugations were detected for the first time in the left middle frontal gyrus, left pre-supplementary motor area and left anterior cingulate cortex.

KEYWORDS:

• Inflection
• regularity/irregularity
• verbs
• inflectional classes
• fMRI

Acknowledgments

We thank Antonietta Canna and Sara Ponticorvo for their help in running the fMRI experiments. We also thank João Verissimo and two anonymous reviewers for their helpful comments on a previous version of the manuscript.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

1 In many languages words occur in a variety of inflected forms (paradigm). Some classes of words called inflectional classes may share a paradigm since they select the same set of inflectional realizations, e.g. verbal conjugations of Romance languages (Aronoff, 1994; Carstairs-McCarthy, 1994). Inflectional class information is a fixed property of words (independent of semantics and phonology) with morphological implications since it specifies the set of inflectional morphemes which are compatible with the word-stems: speakers are allowed to generate all the possible forms included in a word’s paradigm by knowing its inflectional class.

2 Additional verb classes can be found in Italian: there are some verbs (about 0.9%, count performed on Thornton et al., 1997) that do not fit into one of the three main classes, such as porre “to put”. Moreover, some linguists (for instance, Albright, 2002) prefer a different classification where a further distinction is made within the 2nd conjugation between verbs having a tonic accent on the stem like spìngere, “to push”, and verbs without a tonic accent on the stem like cadère, “to fall”. Actually, this classification is not relevant for Italian speakers or Italian L2 learners but it is mostly useful when etymological differences among verbs are considered, since it reflects the origin of the Italian 2nd conjugation verbs from the 2nd and 3rd ancient Latin conjugations.

3 Italian verbs have four finite (or verbal) moods (Indicative, Subjunctive, Conditional and Imperative) and three non-finite (or nominal) moods (Infinitive, Participle, and Gerund). Within the finite moods, Indicative has four absolute tenses: Present, Imperfect, Perfect and Simple Future; Subjunctive has two absolute tenses, Present and Imperfect; Conditional and Imperative have one absolute tense, Present. Finite forms can also be inflected for person and number: each tense for each finite mood has six voices that are combinations of person and number: 1st, 2nd and 3rd singular and 1st, 2nd and 3rd plural. For the non-finite moods, the Participle has two tenses, Present and Past; Infinitive and Gerund have only the Present. These forms are invariable, except for the Participle that, under certain conditions, may take four forms marking gender (masculine and feminine) and number (singular and plural).

4 Count performed on the BDVDB database (Thornton et al., 1997).

5 The experiment conformed to the principles of the Declaration of Helsinki and was approved by the local ethics committee.

6 t-tests performed on values for the controlled variables did not reveal statistical differences across experimental conditions.

7 The experiment conformed to the principles of the Declaration of Helsinki and was approved by the local ethics committee.

8 We used a row of fixation crosses/hash marks rather than a simple fixation cross/hash mark in order to avoid spurious activations in visual brain areas. In fact, switching from a single fixation symbol (a cross or a hash mark) to a string of letters (words) causes the visual stimulation switching from covering a smaller to a bigger portion of the visual field and therefore, expectedly, a higher visual processing load at least for the early visual cortex. In contrast, by using a number of symbols equal to the average number of characters of the presented word strings equalizes the visual processing load between the two visual stimulations. The use of hash marks was chosen to introduce a specific visual cue for the active inflectional task at a fixed delay from the input stimulus (word). In this way, we intended to reduce as much as possible the inter-trial and inter-subject variability in the expected onset of the neural responses associated with the verb inflection. Without such a cue, the inflectional task would have been implicitly triggered by the word stimulus itself with much higher variability in the onset of the verbal responses.

9 Trials in which the voice-key was triggered by accidental noise (external noise, cough, breathing, etc.).

10 In the fMRI task participants almost performed at ceiling level and no response was removed from the analyses.

11 Apparently the Conjugation effect found across regular verbs in behavioural data (2nd conjugation verbs differ from 1st and 3rd conjugation that, in turn, do not differ from one another) does not fully overlap fMRI activation patterns (1st conjugation elicits significantly different activation from 2nd and 3rd conjugation): in our view, the differences in the cognitive demand of the two tasks might explain the inconsistency. The behavioural task requires responses to be given within a temporal window of a few hundred milliseconds. In order to avoid movement artefacts, in the fMRI task, participants can articulate their response after 2200 ms. Reasonably, the difference in timing between the two tasks also entails difference in terms of difficulty: actually, in the fMRI task participants performed at ceiling level while in the behavioural task they produced 11% of responses differing from the target response in many aspects. The fMRI task timing enhances differences between verbs requiring demanding operations of control and monitoring of competing paradigms which might induce interference (2nd and 3rd conjugation verbs) and verbs that can be effortless produced (1st conjugation verbs).

Additional information

Funding

This research was supported by a grant from University of Salerno (FARB, Fondo di Ateneo per la Ricerca di Base, grant n° 300401FRB15DEMARTINO) to Maria De Martino for the research project: “La morfologia flessiva dei verbi nei corpora linguistici, nel lessico mentale e nel cervello”.