Gender Assignment in American Norwegian
Whereas number and definiteness typically vary from one occurrence of the same noun to another (varying singular or plural, definite or indefinite from one occurrence to the other), gender does not vary in the same way and in fact seems to be a fixed property of each individual noun (a given noun always being neuter, for instance). Thus, many formal analyses of the noun phrase treat number and definiteness as categories that are syntactically assigned, whereas gender is taken to be an inherent property of the noun itself (see e.g. Julien 2005). However, there are also analyses where it is assumed that gender is syntactically assigned on a par with number and definiteness (see e.g. Picallo 1991). In our paper, we will use data from American Norwegian (the Norwegian immigrant language in America from c. 1850 onwards) in order to elucidate this theoretical “gender problem”. In particular, we will seek to develop further the analysis of Nygård & Åfarli (2013), arguing that gender is syntactically assigned in a manner similar to number and definiteness.
American Norwegian (Am-No) is particularly interesting regarding the gender problem because this variety of Norwegian shows frequent code-mixing/borrowing of nouns from a language without gender on nouns (English) into a language with a gender system (Norwegian). There are basically two theoretical possibilities for a noun taken from a non-gender system into a gender system: (a) the noun gets a default (“inactive”) gender in virtue of being borrowed (so that all borrowed nouns get the same default gender), or (b) the noun gets a particular (“active”) gender in a systematic way by some assignment rule. The Am-No empirical material collected from the 1930ies (Haugen 1953) until our time (the Am-No material collected in recent years), indicates that English nouns taken into Am-No get different genders in a systematic way, see Hjelde (1996), who finds that of the English nouns borrowed into “Trønder” Am-No, 70.7% are m, 10.5% are f, and 15.7% are n. (Two examples from Haugen 1953: eit nørserikampeni (n) ‘a nursery company’; ein kar (m) ‘a car’.) This suggests that theoretical option (b) is the correct one (see also Hjelde 1996). Thus, we hypothesize that gender is syntactically assigned to the English nouns.
In the second part of our paper, we will inquire into the syntactic base of this assignment rule. We will first argue, based on root theory (Marantz 1997, De Belder 2011) and developing the analysis pursued in Nygård & Åfarli (2013), that any grammatical property of a noun, even its word class category, is acquired syntactically. It follows that all inflectional features, even gender, are acquired syntactically and are not inherent to the noun (root) itself. A structure that captures this intuition is sketched in (1), where n=noun (assigning word class category) and where GEN(der), NUM(ber), and/or DEF(initeness) are hierarchically ordered in the structure.
(1) … [DEF [ NUM [ GEN [ n [ROOT]]]]
We further argue that the essential difference between Norwegian and English is that Norwegian (including Am-No) DPs have the gender assigning GEN functional category/projection, whereas it is absent in English. Therefore any English noun code-mixed or borrowed into Norwegian must be assigned gender. We also argue that an analysis assuming that gender is inherent to the noun itself will be problematic for the code-mixing/borrowing data.
The main challenge for unitary analyses like the one we propose is the problem why gender assignment is not flexible in the same way that number and definiteness assignment is. Based on Nygård’s (2013) analysis of semantic number agreement phenomena, we argue that all inflectional categories (GEN, NUM, DEF) in the functional frame of the noun phrase are unvalued at the start of the derivation, although each “open” category of course makes available a restricted set of possible values. Then, we argue, the inflectional category acquires a particular fixed morphological value by a process called feature construal, where non-linguistic conceptual properties of the lexical item play an important facilitating role. For instance, NUM makes two possible morphological values available (in Norwegian), namely sg and pl, and the actual value of NUM is fixed as either sg or pl, depending of the feature construal of the conceptual “numeric” content of the noun as one or many (both construals are available on the basis of the conceptual content of countable nouns). Thus, the conceptual content of the noun contributes decisively to the choice of the morphological value among the restricted set of possible values made available by the NUM category, and the fixed feature is in turn assigned to the noun as a morphological feature.
Concerning gender, GEN (in Norwegian) makes available three possible values: m, f, n. However, we argue that the actual fixation of the value as either m, f, or n is facilitated in a rule governed way by what we assume is the non-linguistic conceptual gender/class property of the noun, which, unlike the numeric conceptual property of a countable noun, is relatively stable for each individual noun (see Picallo 2008 for a similar analysis based on Romance data). Thus, we get the seeming effect that morphological gender is inherent in the noun, whereas it is in fact the conceptual gender/class property that is inherent, and it is this non-linguistic property that in turn facilitates the fixation of the morphological gender of GEN as either m, f, or n. Nouns in all languages, English included, are associated with various non-linguistic conceptual gender/class properties, but since English lacks the linguistic GEN category, the corresponding conceptual gender/class property is not translated into a linguistic morphological property. However, when an English noun is code-mixed/borrowed into Am-No, the relevant conceptual/class property of the English noun is translated into a linguistic morphological property, visible as gender inflection. We will illustrate this analysis with examples from Am-No, and we will also inquire into the actual gender assignment principles involved.
De Belder, M. 2011: Roots and Affixes. LOT dissertation series, Utrecht University. —Haugen, E. 1953: The Norwegian Language in America (vol. I & II). Philadelphia: University of Pennsylvania Press. — Hjelde, A. 1996: The gender of English nouns in American Norwegian. In: Ureland & Clarkson (eds.) Language Contact across the Atlantic. Tübingen: Max Niemeyer Verlag. — Julien, M. 2005: Nominal Phrases from a Scandinavian Perspective. Amsterdam: Banjamins. — Marantz, A. 1997: No escape from syntax. UPenn Working Papers in Linguistics 4: 201-225. — Nygård, M. 2013: Discourse ellipsis in spontaneously spoken Norwegian: Clausal architecture and licensing conditions. Doctoral dissertation, NTNU, Trondheim. — Nygård, M. & T.A. Åfarli 2013: An Analysis of Variation and Stability in Gender Assignment to Nouns. Paper at 25th Scandinavian Conference of Linguistics (Workshop 10), Reykjavík. — Picallo, M.C. 1991: Nominals and nominalizations in Catalan. Probus 3: 279-316. — Picallo, M.C. 2008: Gender and Number in Romance. Lingue e linguaggio VII: 47-66.