A paradigmatic approach has been applied to derivational morphology in recent studies (Štekauer 2014; Antoniová & Štekauer 2015; Blevins 2016). One of those studies (Rodrigues 2016) proposes the notion of cross-paradigms to describe the intersection of process-organised with affix-organised paradigms. This paper brings psycholinguistic evidence to the mental status of cross-paradigms, based on experimental data obtained from native Portuguese speakers. Cross-paradigms reveal the dynamic nature of the mental lexicon, as Elman (2011) and Libben (2014; 2015) have claimed.
Our aim is to demonstrate with Portuguese word-formation data collected from corpora (Linguateca, Corpus de Referência do Português Contemporâneo and Corpus do Português) and based on experiments with Portuguese native speakers that derivational paradigms are mental patterns dynamically organized around more than one axis in what we call cross-paradigms. Cross-paradigms are structured by affixes which may put different base-organized paradigms into interface.
This hypothesis is supported by recent psycholinguistic approaches to the mental lexicon such as Libben (2014) (cf. Libben’s concepts on morphological transcendence and morphological superstates), according to whom the mental lexicon is not as an «inert knowledge store, [but] as a dynamic cognitive system that allows for lexical activity.» (Libben 2014: 209). Our hypothesis is also founded on linguistic works such as Corbin (1987), which, as demonstrated by Booij (2007), presents a paradigmatic perspective on word-formation, and on Blevins (2016).
Following Štekauer (2014: 359), we consider derivational paradigms as «based on formal realization of a cognitive category by an affixation process.».
According to Pounder (2000), different materials can organize each paradigm. It is not just a specific morpheme that functions as the axis of a certain paradigm. The axis may correspond also to the word-class, to semantic rules, and to other features labelled under ‘lexical paradigm’ by Pounder. Our work sticks to two main different paradigm relationships: the lexeme-base-class-organized paradigm and the affix-organized one.
Giving examples from Portuguese, a lexeme-base-class organized paradigm is illustrated by deverbal nouns with different suffixes such as avaliação ‘evaluation’, matança ‘slaughter’, congelamento ‘freezing’, aterragem ‘landing’ and soldadura ‘soldering’. The axis of this paradigm corresponds to the base lexeme the nouns correlate with, which is a verb (avaliar ‘to evaluate’, matar ‘to kill’, congelar ‘to freeze’, aterrar ‘to land’, soldar ‘to solder’).
An affix-organized paradigm is illustrated by nouns such as medievalismo ‘medievalism’, espiritualismo ‘spiritualism’, luteranismo ‘Lutheranism’, newtonianismo ‘Newtonianism’ and figurativismo ‘figurativism’. The axis of this paradigm is the suffix -ism(o).
Models that propose separated paradigms like those collide with empirical data. In Table 1, we show three suffixes (-ism(o), -eir(a) and -agem) that work with different lexeme-base classes. We may exemplify this assumption by means of the suffix -ism(o). This suffix may form nouns correlated with lexeme classes other than adjectives: correlated with verbs (bisbilhotar ‘to gossip’bisbilhotismo ‘habit of gossiping’) and correlated with nouns (sigilo ‘stealth’sigilismo ‘secretiveness’). The fact that nouns with the suffix -ismo correlate with verbs, nouns and adjectives creates an interface between the three lexeme-base-class paradigms.
