Morphological Productivity Measurement: Exploring Qualitative versus Quantitative Approaches
2013; Routledge; Volume: 94; Issue: 4 Linguagem: Inglês
10.1080/0013838x.2013.780823
ISSN1744-4217
Autores Tópico(s)Syntax, Semantics, Linguistic Variation
ResumoAbstractMorphological productivity was originally described as the property of word-formation processes whereby new words are created to satisfy a naming need, a definition later enhanced with Corbin's classical distinction between availability and profitability. Despite the importance of these two concepts, they have been traditionally overlooked in the design and discussion of productivity computations. Moreover, no thorough exploration has been made of the implications they have in existing models of productivity measurement. This paper examines the treatment given to this notional pair in the mainstream productivity methods by, first, examining their theoretical background and, second, discussing their practical implications in their application to a BNC-derived corpus. Subsequently, a number of suggestions are provided, and an alternative unifying method is proposed which is able to provide insights into both availability and profitability in the computations of morphological productivity of word-formation processes. Notes1Schultink, 113.2E.g. Hockett, 15–16; Marchand, 139–40, 263.3In his view, because some processes have more bases to which they attach than others, it seems more accurate to measure each process individually by considering its own scope of application (cf. Berschin, 44–5).4Corbin, 177.5Danielle Corbin's distinction contemplates, in fact, three sides to morphological productivity: availability, profitability and regularity. Today, regularity is not analysed on a par with availability and profitability, even if it is still perceived as a property of morphological processes (cf. Bauer, Morphological Productivity, 54–6). The reader is referred to the following references for further details: Kastovsky, 586; Spencer, 48–9; Plag, Morphological Productivity, 34.6Zandvoort, 44; and Lyons, 76, respectively.7Sproat and Shih; Baayen, "Probabilistic Approaches"; Nishimoto.8The British National Corpus, version 2.9Aronoff, Word Formation, 36.10Baayen and Lieber, 802–6.11Cf. Aronoff, "Potential Words"; Plag, Morphological Productivity, 24; Bauer, Morphological Productivity, 145–6.12Bauer and Huddleston, 1666.13The labels used for the syntactic categories are AJ for adjective, AV for adverb, DT for determiner, NN for noun, OR for ordinal and VB for verb. All examples come from the study corpus unless otherwise stated.14OED.15Affixes are sorted out from more to less productive hereafter unless otherwise stated.16Cf. Bauer, Morphological Productivity, 205, for a claim of the unavailability of the suffix.17Dieter Kastovsky exemplifies this by arguing that not every word ending in -ion or -ive is synchronically relevant for English word-formation, and that sometimes the units existed individually in French. This is the case of the series deceive > deception > deceptive, where no word-formation has taken place in English, and which rules out any base > derivative relationship (cf. Kastovsky, 589).18Bauer, Morphological Productivity, 144.19Baayen, "On Frequency"; Plag, Morphological Productivity; Plag, Dalton-Puffer and Baayen; Hay, Causes and Consequences.20Cf. Schultink, 113.21Hapax legomena are words that occur only once in a corpus, that is where N = 1 (cf. Baayen and Lieber, 809).22Aronoff, "Potential Words"; McQueen and Cutler.23Baayen, "On Frequency," 183.24Given the different frequency ranges of the two word-lists used here (61,847 to 10 vs. 6,187,267 to 1, see section 2), the N values of the former were multiplied by 10,000 to make the frequencies compatible with Kilgarriff's word-list (cf. Leech, Rayson, and Wilson, for further details on this point).25Cf. Bauer, Morphological Productivity, 148–9; Gaeta and Ricca.26Baayen, "On Frequency," 205.27Ibid.28Baayen and Lieber, 809.29Bauer, Morphological Productivity, 148.30Lüdeling, Evert, and Heid, 3.31Hay, "Lexical Frequency"; Hay, Causes and Consequences; Hay and Baayen.32Hay, Causes and Consequences, 160–80.33Cf. Cowie; Bauer, "Productivity: Theories," 328.34Schultink, 113; Plag, Morphological Productivity, 22; Bauer, Morphological Productivity, 14.35Booij, 121; Spencer, 48; Dressler and Ladányi, 119.36Štekauer, Onomasiological Theory; Štekauer, "Fundamental Principles"; Štekauer, "Onomasiological Approach"; Štekauer et al.37Štekauer, "Fundamental Principles," 5.38In this theory, the term naming unit is equivalent to word, lexeme or lexical unit (cf. Štekauer, "Onomasiological Approach"; Štekauer et al.).39Chomsky; Baayen and Lieber.40Productivity has been claimed to include rule-governed coinages, while creativity takes in any other kind of output, for example literary texts, playful formations or terminological units (cf. Dressler and Ladányi, 105; Bauer, Morphological Productivity, 62–71). However, to the best of my knowledge, no study has clearly separated them in practical terms.41Cf. Štekauer, Onomasiological Theory; Fernández-Domínguez, Díaz-Negrillo, and Štekauer; Štekauer, "Fundamental Principles"; Štekauer et al. In these studies, the study sample never exceeds 350 naming units. Note the sharp contrast between this figure and R. Harald Baayen's habitual eighteen-million-word corpus (Baayen and Lieber, 801–5).42Fernández-Domínguez, 147–69.43Kastovsky; Štekauer, "Fundamental Principles."44See fn. 40.45Even if this model is here applied only to affixation for the purposes of this article, its scope makes it possible to apply it to any word-formation process. Cf. Fernández-Domínguez for an application of the model to N + N compounding.46Cf. Aronoff, "Potential Words"; Bauer, Morphological Productivity.47The use of V and N for productivity measurement was declined in 2.1.1 given their objectionable use as straight counts. However, it should be emphasised that such a rejection refers to the use of type and token frequency in isolation, that is as direct computations of productivity. The present proposal, by contrast, is to modulate such figures in order to obtain a sign of what they may indicate about a given morphological process, regardless of the implications that V and N may individually carry.48Note that π is an approximation towards the profitability of word-formation processes and not towards their productivity in global terms. Precisely for this reason, it can be applied only to available processes (cf. Fernández-Domínguez, 154–6).49See fn. 17.50Hay and Baayen, 101; Plag, Word-Formation in English, 205–6.51Cf. Corbin, 177; Bauer, Introducing Linguistic Morphology, 61.52Bogdan Szymanek explains that there are, in fact, two homophonous -nik suffixes, one of which stopped being productive in the sixties, the other being still productive in Yiddish-English. However, despite its current activity, the latest record of this affix in the OED was attested in 1975, which brings problems when providing an authoritative source for the present availability of the affix (cf. Bauer, English Word-Formation, 255–66).53Cf. Marchand, 293; Bauer, Morphological Productivity, 62–71; Bauer, "Productivity: Theories," 329–30.54Štekauer, Onomasiological Theory, 82–5.55Aronoff, 30; Bauer, Morphological Productivity, 138; Plag, Word-Formation in English, 67–8.56Saussure, 221–37; Fernández-Domínguez, 63–5.57Plag, Morphological Productivity, 12.
Referência(s)