Terminology for verbal derivation

[Daniel Everett]

Martin,

As I said in my initial response, I agree that the mnemonic use of these labels can be useful. The same holds for all linguistic terms (outside of particular theories). And even for the IPA - to write a word using the  IPA is to simply limit the range of its phonetics, not to describe or analyze it. a [p] in English is not the same as a [p] in Pirahã, for example, since the lips are drawn flatter across their entire length in the latter than in the former. 

But too often we see terms used in lieu of analysis, in grammars and in journal articles. This is a much more severe problem in some theories than in others. 

It is perhaps worse in syntax. Someone might say that "this is not a passive because there is no by-phrase" or "this is a passive because there is a reduction in transitivity from the active." I think both of these statements are the wrong way to go about the job. Describe the construction by giving the full range of observed possibilities (participating verbs, kinds of arguments, use of other phrases, etc). Then you can say something like "it shares core features with the passive" or some such. 

We need to use mnemonic devices, I agree. And Bernard's 1976 proposal is a reasonable one. There are bound to be similarities between languages because of the functions that are performed, so we need a way to cross-reference similar solutions to similar problems. But we need a great deal of caution as we do. These labels are just ways of beginning discussion between linguists. As you say, they should not be considered analyses.

Like any words. What is a "dog?" The range of referents for this word will vary by region of the world. There will be overlap with English but no precise alignment. Different kinds of terms,  less observable or concrete than "dog"s will overlap less. So we should not be afraid to refer to things as "dogs." So long as we add a list of referents of the term or an explanation of the system behind its use in different languages.


Dan


On Jul 17, 2012, at 2:23 AM, Martin Haspelmath wrote:

> The "awkwardness" disappears if we conceptualize our grammatical terms 
> as mnemonic labels, rather than as analytical decisions.
> 
> Thus, we should not say things like "I analyze this form as an article", 
> or "I analyze this form as an applicative". Analysis does not consist in 
> attaching a pre-established label to a form, but in "copious 
> description" (as Dan noted).
> 
> But if we just number the forms that we find, or call them by their 
> shape, e.g. "the -ia form", then it's very difficult to talk about 
> language structure (as Michel noted).
> 
> So the best solution, to my mind, is to use grammatical terms adopted 
> from another language (from Latin, or from Nahuatl, or from English) as 
> mnemonic labels, and to capitalize them to show that they are 
> language-specific items ("the Arabic Article", "the Turkish Genitive 
> case", etc.). This was proposed by Bernard Comrie in 1976 and has proved 
> to work well in typology (see the discussion in my 2010 paper on 
> comparative concepts and descriptive categories, in Language).
> 
> Greetings,
> Martin
> 
> Dan Everett wrote:
>> I agree that it is an awkward business. But I am aware of the implications. It is OK to talk about "genitive," "article," "applicative," "voiceless bilabial," and so on as long as we are aware that none are precise and all must be accompanied by copious description - to the point that they are ways of cutting down the solution space at best, harmful at worst. 
>> 
>> Dan
>> 
>> Sent from my iPhone
>> 
>> On Jul 16, 2012, at 17:56, "Michel LAUNEY" <michel.launey at ird.fr> wrote:
>> 
>> 
>>> All linguists will agree that cross-linguistic variation is great, and that our terminology is therefore unavoidably awkward, but Dan Everett's reluctance may lead to still more awkward consequences: either coining a different terminology for each individual language, or giving up any terminology.
>>> For instance, this would forbid us to say that English, French and Arabic all have definite articles, that, say, Latin, Russian, Modern Greek and Turkish all have a genitive case, that English, French and Spanish all have reflexive forms etc. etc., because what we call so have different uses in each language.
>>> I wonder if being wary about "applicative" (and, I suppose, much less about "article", "genitive", "reflexive" and so on) does not come from the fact that what "applicative" (or "benefactive", "prepositional form", "objective version" etc.) refers to is a relatively "novel" discovery in the history of linguistics, so that linguists are a bit touchier about the specificities of the language or language family in which they find the phenomenon, while much more tolerant about the awkwardness (call it polysemy or even ambiguity if you like) of more formidably traditional terms for parts of speech or grammatical categories.
>>> Best
>>> Michel Launey
>>> 
>>> On Mon, 16 Jul 2012 15:51:09 -0400
>>> Daniel Everett <dan at daneverett.org> wrote:
>>> 
>>>> It is possible that 'applicative' is the best term here, depending on local linguistic traditions.
>>>> But as Sally and I (and different co-authors) showed in a series of papers on Salish the range of ways to modify/signal modification of various manifestations of transitivity and valence go beyond currently available terminology. I don't mind terms like 'applicative' as mnemonic devices in limited contexts, e.g. specific language families, but I don't like them when they are used as cross-linguistic standards. I don't find rigid use of terms all that useful. The variation is too great in most cases, especially when looked at more carefully.
>>>> -- Dan
>>>> On Jul 16, 2012, at 3:44 PM, Michel LAUNEY wrote:
>>>> 
>>>>> Hi,
>>>>> "Applicative" seems to me, definitely, the best term.
>>>>> To my knowledge, it was first coined in 1595 by Antonio del Rincon, who in his "Arte Mexicana" had "discovered" this phenomenon in Nahuatl. There is a long tradition of use of this term in grammars of Nahuatl and other Middle American languages.
>>>>> In the Bantu linguistic tradition, it is also used, but more often you will find "prepositional form of the verb" (which is strange, because precisely the added argument NP occurs with no preposition).
>>>>> I find in a Georgian grammar (Tschenkeli "Einführung in die georgische Sprache") "Objektive Version", which seems to me also unsatisfactory.
>>>>> Best
>>>>> Michel Launey
>>>>> On Mon, 16 Jul 2012 07:54:10 +0100
>>>>> Lachlan Mackenzie <lachlan_mackenzie at hotmail.com> wrote:
>>>>> 
>>>>>> Hi, John,
>>>>>> To me it seems like 'applicative' might be the word you're looking for. One applicative form can cover various meanings, in the way you describe for Bari -kindya.
>>>>>> Cf. David A. Peterson (2007). Applicative Constructions. OUP.
>>>>>> Best wishes,
>>>>>> Lachlan Mackenzie
>>>>>> 
>>>>>>> To: funknet at mailman.rice.edu> Date: Mon, 16 Jul 2012 08:40:26 +0300
>>>>>>> From: john at research.haifa.ac.il
>>>>>>> Subject: [FUNKNET] Terminology for verbal derivation
>>>>>>> Dear Funknetters, I'm looking for a term to use to refer to a form
>>>>>>> for deriving verbs in Bari (-kindya) which seems to generally add an
>>>>>>> argument to the verb, but the argument can be any one of a variety of
>>>>>>> types--it can be an indirect object, a directional particle, just
>>>>>>> about anything it seems (for example, when added to the root meaning
>>>>>>> 'old age', it can take as an argument a place, with the meaning 'to live
>>>>>>> to an old age while living continuously at that place', or a
>>>>>>> nominalized form of a verb referring to an occupation, with the
>>>>>>> meaning 'to live to an old age while continuing to work at that
>>>>>>> occupation'). Do you have any ideas what term I might use to refer to
>>>>>>> this form of the verb? I was initially going to call it the
>>>>>>> 'Benefactive' because it's often used to add an indirect object (e.g.
>>>>>>> 'close a door for someone') but when I looked at all of the usages of
>>>>>>> this form it became clear that this is really a pretty small minority of
>>>>>>> its uses. Any ideas? Thanks, John 
>>>>>>> 
>>>>>> 
>>>>>> 
>>> 
>> 
>> 
>> 
>> 
>