[Lingtyp] grammaticalized v grammaticized

Alex Francois francois at vjf.cnrs.fr
Wed Jun 12 12:35:59 EDT 2019

dear Prof. Lehmann, dear all,

Christian Lehmann's message prompted some thoughts while I was in the tram.

> *Nobody has yet proposed to distinguish between synchronic and diachronic
I'm not sure about that.  Some processes of assimilation are productive,
subject to active rules in the synchrony of a language [*wug * > *wug-s*];
in other cases, assimilation is really a diachronic process, which has no
reality in modern languages [Lat. **ad-simulātio* >  **assimulātio*  > Eng.
I think it does make sense to distinguish between synchronic and diachronic
assimilation.  The same could be argued about the other examples cited
(diphthongation, univerbation, metaphora).

About *grammatic(al)ization*:
> *I would advise against providing different terms for the concepts
‚synchronic grammaticalization‘ and ‚diachronic grammaticalization‘.*
I do think the distinction makes sense here, and for another important
reason:  namely, the two phenomena discussed in this week's debate are
actually quite distinct.

Let me quote again Bill Palmer's very clear formulation:

I use *grammaticized *to refer to a synchronic situation, and *grammaticalized
> *to refer to a diachronic process. For example, I would say that the
> category of auditory evidentiality ("I heard [X happen]") is grammaticized
> in language X, meaning that the category is expressed in the language by a
> grammatical form; and I would say that the verb 'hear' has grammaticalized
> as an evidential marker in language X, meaning that a form with a lexical
> meaning has developed into a grammatical marker of some kind.

I think that Bill's two categories differ not only by the trait *synchronic*
 vs. *diachronic*, but also by other properties;  in effect they are two
separate concepts.

Although I've never personally used the forms "grammaticization" or
"grammatization", I will follow Bill's usage (at least for the sake of the
current discussion), and distinguish between two terms:

   1. "*grammaticization*" is a property of a system.
   It refers to the fact that a given grammatical system (a language L)
   treats a given meaning M as an emic category, using formal structures
   (morphs, morphemes, constructions) that are contrastive.

   2. "*grammaticalization*" is a property of a form.
   It refers to a process of linguistic change, whereby a form F, usually
   understood to be initially lexical, acquires a meaning M that is
   grammatical.  That meaning M may exist already in the system, or it may be
   created precisely as a result of this process of grammaticalization.

[*a note on terminology*:
I reckon that "grammaticalization" in the sense #2 is well established in
our discipline, and should not be changed.  "Grammaticization" in the sense
#1 is less entrenched, and it is not too late for us to find a better
term.  Grammatify? grammatize? grammemicize (cf. phonemicize) ?  or just
*emicize* ? hm.]

Crucially, #1 and #2 are orthogonal phenomena;  they cannot simply be
analyzed as two facets (one synchronic, one diachronic) of the same thing.

Indeed, #1 grammaticization itself can be understood as compatible with a
synchronic and a diachronic interpretations, distinct from #2

   - #1 grammaticization viewed synchronically: *language L treats meaning
   M as emic*.
   - Classical Latin grammaticized the TAM category of future:
      *scrībet*  'he will write'  (opp. Indicative present *scrībit*,
      Subjunctive *scrībat*).
      - Homeric Greek grammaticized the Dual: τὼ χεῖρε "(both) hands".
      - Hiw (Vanuatu) grammaticizes the subjunctive — i.e. it "has" a
      subjunctive category, contrary to its neighbours.

   - #1 grammaticization viewed diachronically: *language L acquires an
   emic category M which it didn't have before*.
   This takes place via language change [but not necessarily via #2

   - e.g. 17th century French has been claimed (Lancelot & Arnauld 1660,
      cited in Comrie 1985:93) to have acquired a semantic contrast
between *hodiernal
      past* /j'ai écrit/ 'I wrote [earlier today]' vs. *prehodiernal
past* /j'écrivis/
      'I wrote [earlier than today]'. This could be formulated by saying that
      French, historically, began *grammaticizing *a category of hodiernal
      past (as in, say, Yeyi, a Bantu lg of Namibia – Seidel 2008), which
      it didn't use to grammaticize before — and which it later lost.

      - Other example:
      While PIE used to grammaticize the two cases Instrumental and
      Ablative separately, Classical Latin merged them;  as a result,
for a while
      it grammaticized an emic category of *instrumental–ablative* [= the
      classical Ablativus: *gladiō* 'with a sword' ~ *domō* 'from the

      Later on, that emic category ceased to be grammaticized, as Late
      Latin (Romance) developed a contrast between a proper ablative on the one
      hand [**de suo domo* 'from his house'] vs. a new category of
      instrumental–comitative [**cum suo gladio* 'with his sword' / **cum
      suo patre *'with his father'].

These processes of semantic change would not be accurately described as
cases of #2 "grammaticalization", if the latter is understood as the change
of a form from lexical to grammatical.  The emergence and/or loss of
grammaticized (emic) categories took place here through the semantic
restructuring of existing grammatical categories, with no lexical input.

As for #2 *grammaticalization*, it could also be argued to be compatible
with a synchronic and a diachronic interpretation:

   - synchronic:  *a form F is used, synchronically, both as a lexical item
   and as a grammatical device*.
   e.g. in Mwotlap (Vanuatu), the form *tēy* is a lexical verb meaning
   'hold, have in hand', and it has also grammaticalized as an applicative.
   Both meanings are active in synchrony.  In the same language, *van* is
   both a verb 'go' and an allotropic directional ("thither").

   French also has a form *devoir*, which is both a lexical verb (*je lui
   dois 30 euros* "I *owe* her 30 €") and a grammatical device coding for
   deontic or epistemic modality (*il doit être 3h* "it must be 3 o'clock").

   Synchronic grammaticalization is a type of polysemy, and more exactly,
   of heterosemy (cf. Lichtenberk 1991; Enfield 2006).

   - diachronic:  *a form F, at some point, loses its lexical meaning, and
   survives as a grammatical morpheme*.
   e.g. Lat. *habēre* 'hold, have' grammaticalized as an auxiliary for
   perfect tense [*habeo scriptum* 'I have written'], while still keeping
   for a while its lexical meaning — resulting in heterosemy (within Latin).
   Modern Spanish then lost the lexical meaning of *haber*, and keeps this
   form only as a grammatical device.

Finally, on the relations between #1 and #2:  these are orthogonal concepts.

I'll try to put my thoughts in order with a table:
[image: image.png]
Notes about this table:
#2 grammaticalization comes with diachronic #1 grammaticization, *iff* the
change from Lexical to Grammatical results in the emergence of a
grammatical category that did not exist previously in the system.

e.g. Classical Latin used to not grammaticize the conditional (=it didn't
have it as an emic category).
At some point, Late Latin went through (#2) grammaticalization **scrībere
habēbat* 'he had to write' >> Ital. *scriver-ebbe* 'he would write'.  This
instance of #2 grammaticalization was also, in this particular case, an
instance of diachronic grammaticization (the creation of a new grammatical

Conversely, the (#2) grammaticalization **scrībere habet* 'he has to write'
>> Ital. *scriver-à* 'he will write' did not entail a change in #1
grammaticization, since the earlier system already had a category of
future.  This was just a case of morphological renovation (via lexical
input and grammaticalization) of an existing category.  A diachronic
process of grammaticalization, with no change in grammaticization.


Alex François

LaTTiCe <http://www.lattice.cnrs.fr/en/alexandre-francois/> — CNRS–
–Sorbonne nouvelle
Australian National University
Academia page <https://cnrs.academia.edu/AlexFran%C3%A7ois> – Personal
homepage <http://alex.francois.online.fr/>


On Tue, 11 Jun 2019 at 11:09, Christian Lehmann <
christian.lehmann at uni-erfurt.de> wrote:

> The discussion has brought up several distinct meanings that can be
> associated with such terms as *grammat(ic(al))ization*. Several of the
> concepts involved are doubtless useful in linguistics and would suit such a
> term. I would advise against providing different terms for the concepts
> ‚synchronic grammaticalization‘ and ‚diachronic grammaticalization‘. I
> distinguish between the formation of a concept and providing a term for it.
> There is apparently no methodological principle that would allow or exclude
> the formation of concepts of just anything. Certainly it may be useful to
> distinguish between ‚grammaticalization viewed as manifested in synchrony‘
> and ‚grammaticalization viewed as manifested in diachrony‘. However, one
> has to keep in mind that synchrony and diachrony are not two different
> spheres of the object of linguistics, but two alternate perspectives on one
> object. Thus, there are no such things as ‚synchronic grammaticalization‘
> as a process distinct from ‚diachronic grammaticalization‘. The same is
> true of countless other linguistic processes. Nobody has yet proposed to
> distinguish between synchronic and diachronic assimilation, synchronic and
> diachronic diphthongation, synchronic and diachronic univerbation,
> synchronic and diachronic metaphora and so on ad nauseam. Descriptive and
> historical grammarians have simply assumed that there is, in each of these
> cases, only one such process which manifests itself in the perspective
> taken by them; and rightly so.
> So again, one may, of course, view grammaticalization either in a
> synchronic or in a diachronic perspective. It is, however, methodologically
> dangerous to provide different terms for such constructs, because a
> construct provided with a (handy) term has a strong tendency to be
> hypostatized to an entity existing independently of our approach. Witness
> the countless definitions found in the literature according to which
> grammaticalization is allegedly a diachronic (or even worse, a historical)
> process. Sorry for sounding dogmatic about this; but our theory is going to
> make progress only if we get the methodology right.
> Positive balance: Let’s reserve the variants of *grammat(ic(al))ization* for
> some of the other concepts brought up in the discussion.
> --
> Prof. em. Dr. Christian Lehmann
> Rudolfstr. 4
> 99092 Erfurt
> Deutschland
> Tel.: +49/361/2113417
> E-Post: christianw_lehmann at arcor.de
> Web: https://www.christianlehmann.eu
> _______________________________________________
> Lingtyp mailing list
> Lingtyp at listserv.linguistlist.org
> http://listserv.linguistlist.org/mailman/listinfo/lingtyp
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20190612/f9cda162/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 103633 bytes
Desc: not available
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20190612/f9cda162/attachment-0002.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Francois_2019-06-12_notes-on-Grammaticalization.png
Type: image/png
Size: 60729 bytes
Desc: not available
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20190612/f9cda162/attachment-0003.png>

More information about the Lingtyp mailing list