In today’s edition of curiosities from the garden of Basque grammar comes a brief consideration of dictionary citation forms, specifically nominals.
Let’s start by considering the word for ‘daughter’ and the phrase for ‘three daughters’. Look these two up and you’ll get the following: ‘daughter’ = alaba ‘three daughters’ = hiru alaba Great. Now let’s try the word for ‘horse’, which I’ll tell you is zaldia. So, you’d assume the phrase for ‘three horses’ would be hiru zaldia, but actually it’s hiru zaldi. Why? As Rudolf de Rijk writes in his seminal work Standard Basque: a progressive grammar, “When asked what the Basque word is for ‘fire’, any Basque speaker – unless he happens to a Souletin – will answer “sua.” Likewise for ‘black’: “beltza.” The normal or unmarked form of a noun or adjective is for him the definite form, the form carrying the article -a. Consequently, this article occurs with a much greater than the definite article in English” (page 35). Similarly, R.L. Trask’s The History of Basque mentions straight from the get-go (page xx), that nouns and adjectives are cited as without the definite article suffix e.g. gizon ‘man’ and berri ‘new’ rather than gizona ‘(the) man’ or berria ‘(the) new’, even though “Most native speakers, however, differ in that they cite such words with the article attached”. In my efforts to teach myself Basque, this practice of citing nominals with the definite suffix attached is frustrating, because it obscures the underlying structure of the root. That is, some roots end in -a and the definite (absolutive) singular does not add an additional -a, thus alaba is both indefinite singular and definite singular, whereas gizona is definite singular, but indefinite singular is gizon. Google translate, it seems, is not terrible helpful to find the root structure. For example, as of this writing, ‘black’ is translated as beltza, rather than beltz, ‘wine’ as ardoa or ardo, but ‘man’ is only gizon. The trick that I only recently discovered is to look up an entire phrase with numeral, since numerals are usually used with the bare stem (... I guess Slavic missed the memo on this!). Using this method, ‘three horses’ is translated as hiru zaldi, but ‘three truths’ is hiru egia. Thus, the latter therefore has an -a in the stem, the former does not. All of this, however, got me thinking about what counts as a dictionary citation form and why we use the ones we do. In English, it would be odd to cite the man as the dictionary citation form, but we do cite verbs, excluding modals, with the to infinitival particle e.g. to see, whereas in closely related German and Dutch the infinitive is cited without the infinitival particle e.g. sehen, not *zu sehen and zien, not *te zien. Perhaps this is because English to-infinitives are more frequent than their German and Dutch equivalents, although this was less obvious to me when I thought about it more deeply, but I suppose the use of to-infinitive complements with want and (usually) need, but never with wollen and (in colloquially speech) not with brauchen counts for something. In other languages, citation forms for verbs differ. For example, in Bulgarian the first person singular present indicative is used as citation form, since the language lacks infinitives. Thus, the verb for ‘to write’ is cited as пиша ‘I write’ and ‘to eat’ is ям rather than something like писать or есть as found in Russian. Infinitive loss sets Bulgarian apart from other Slavic languages and is a property of Balkan languages more broadly, including Modern Greek, Romanian, Albanian and others, which have all limited the use of infinitives. The classical languages of Latin and Classical Greek are also usually cited under their first person singular forms rather than their infinitives, though this does not always disambiguate class membership. For example, in Latin portō ‘I carry’ and scrībō ‘I write’ belong to two different verbal classes (I and III). For this reason, Latin verbs are usually cited in their four principal parts (1st singular present tense, infinitive, 1st singular perfect, supine) e.g. portō-portāre-portāvī-portātum or scrībō-scrībere-scrīpsī-scrīptum. In Sanskrit, verbs are commonly in their stem form, which never actually occur as distinct words. For example, ‘to strike, beat’ is √han-, which has a class 2 Parasmaipada (active) conjugation pattern. (Classical) Sanskrit textbooks, grammars and dictionaries usually provide this stem with the checkmark notation for stem (‘√’) and then the class number (there are 10 verbal classes for the present system!) and the notation for active or middle voice. The fact that the choice of stem reflects the Sanskrit grammarian analysis of stem and not a modern Indo-Europeanist one can often be seen. For example, √han- has a third person paradigm like the following hanti ‘s/he strikes’, hataḥ ‘they two strike’, ghnanti ‘they strike’. Thus, the verbal root could probably be constructed as √hn- (=hØn), with full grade in the present active singular and zero grade elsewhere. Thus, *hn̥-taḥ > hataḥ (with syllabic nasal regularly becoming short a) and *(g)hn-anti > ghnanti (with consonantal nasal surviving). Logically, one would expect that citation forms would have one of the two properties:
It seems that the Basque practice of citing nominals with -a is preferring property 1 over property 2. This is especially clear for adjectives, which frequently occur with the -a suffix, since attributive adjectives follow nouns (and only NP-final elements get endings) and predicative adjectives require the definite form. Bulgarian also emphasizes the importance of property 1. After all, you can’t use the infinitive as the citation form if it no longer exists in the language! Sanskrit’s use of an unattested verbal stem probably reflects the fact that that language has one of the most ancient traditions of grammatical analysis.
0 Comments
Your comment will be posted after it is approved.
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. Archives
August 2023
Categories |