LINGUIST List 6.300

Sun 26 Feb 1995

Sum: Function Words

Editor for this issue: <>


  1. Kai von Fintel, Sum: Function Words

Message 1: Sum: Function Words

Date: Fri, 24 Feb 95 15:45:44 EDSum: Function Words
From: Kai von Fintel <fintelMIT.EDU>
Subject: Sum: Function Words

I. The Question and its Motivation

Three weeks ago or so, I asked LINGUIST readers to help me find examples
of function words with unusually specific meanings. This was meant to
elicit examples that I could mention in a paper called "The Formal
Semantics of Grammaticalization", where I discuss what happens to the
meaning of a content morpheme that develops into a function morpheme. I
argue there, developing work by other semanticists/logicians, that

(1) Function morphemes do have meanings (contrary to a widespread
 assumption). Just consider that the meanings of quantifiers, tense,
 aspect, modals, etc. are the bread and butter of us working

(2) There is a class of "logical" meanings that have certain formal
 properties (permutation-invariance, high types).

(3) Unfortunately, there is no perfect correlation between logical
 meanings and functional morphemes. There are lexical/content
 morphemes with (almost) logical meanings (adjectives "same", "mere",
 "alleged", verbs "deny", "believe", nouns "majority", etc.). And
 there might be functional morphemes with non-logical meanings (hence
 the LINGUIST query).

(4) Thus there is not much that we can say about what happens to the
 meaning of a morpheme that is becoming grammaticalized. Although I
 make some presumably doomed efforts.

The finished paper (which is based on a talk I gave at the NELS 25
Workshop on Language Change) will appear in the NELS 25 Proceedings (to
be available from the GLSA at UMass Amherst,
The paper is downloadable by anonymous ftp from the following URL:

I would welcome any comments.

II. Some of the Responses:


) The reference [to the pejorative pronouns "you shit"]
) is Samuel Elbert's grammar of
) Rennellese/Bellona, "Echo of a Culture: A Grammar
) of Rennell and Bellona", UHawaii Press 1988
) [Oceanic Linguistics Special Publication No. 22].
) A few interesting cases of
) grammaticalization of a somewhat relevant
) type do exist. In Micronesian languages
) there are a set of verbal affixes for
) directionality (up, down, towards speaker,
) towards hearer, towards some third
) party [the usual deixis system]) which
) also include "toward the open ocean"
) and "toward the lagoon"; whereas (of course)
) if you want to say "toward John's house"
) (or some other nonce collocation)
) you need to use a PP.
) Similar, but probably not similar enough,
) is the bizarre use of the cardinal directionals
) in Icelandic documented in a classic paper
) by Einar Haugen (I can dig up the reference
) if you really want it). It seems like, being
) predominantly coastal inhabitants, and having
) a rather uneven fjorded coastline, to go
) to a city which was north of you actually
) involved heading out in a southerly
) direction (all that time). (Or, if you
) lived on the North Coast, going South
) frequently involved your walking North
) or Northeast...). The system ended up totally
) screwed up, as I recall (it's been some
) time since I read the article -- Hoski
) Thrainsson at Harvard would probably know
) the relevant facts []),
) when the cardinal directions were grammaticalized
) with inverse force from their original
) semantics with verbs of motion (or some
) such thing).
) The "river names" case is supposed to Salishan.
) Maybe Sally Thomason will respond to your
) Linguist posting, otherwise you might write
) to her ( Although
) the Salish speakers she's actually working
) with are land-bound, she's done some historical
) Salishan and probably knows the basic facts.
) The only other case I can think may not be
) all that relevant, either, I guess, but it's
) kind of interesting nonetheless. In Ho-Min
) Sohn's "Woleaian Reference Grammar", I suspect
) in the discussion of noun incorporation
) (but possibly elsewhere), he states that
) bare N objects obligatorily incorporate
) (so 'I eat fish' is ungrammatical, one has
) to say 'I fish-eat' -- this is generally
) true with Micronesian transitive clauses;
) I've written about it's history). Definite
) NP objects cannot incorporate (*I the-fish eat).
) Nor can N's that are modified by anything
) (so 'I eat big fish' is fine, in spite of
) the fact that 'I eat fish' is garbage -- clearly
) a structural constraint: only heads can
) incorporate). Anyway, "doctor" cannot
) incorporate under any circumstances, because
) there's only ever been ONE doctor on Woleai!
) So it's inherently specific, as it were.
) [Woleai is an atoll, population ca.250.]
) Like I said, maybe not directly relevant,
) but kinda cute...

David Gil ( wrote:

) I can think of so many examples of what you're looking for that
) I suspect there's some terminological confusion.
) Classifiers. They're about as idiosyncratic and as contentful
) as you'd like -- but their uses are clearly functional. Arguably
) so when in "numeral classifier" position, more clearly so when
) functioning as nominalizers (eg. CLF John saw, meaning "the one
) John saw"), ligatures/relativizers (eg. movie CLF John saw,
) meaning "the movie which John saw", or articles (eg. CLF movie,
) meaning "the movie"). (I'm presently working on the syntax
) and semantics of these constructions in SE Asian languages.)
) Then, pronouns. I don't know of any "you shit" examples, but
) in SE Asian languages pronouns come with all sorts of idiosyncratic
) and culture-bound "honorific" content.
) I could go on ...

Hala'sz Sa'ndor ( wrote:

) Are the old Germanic prepozitions enough for you: "benorth", "beeast",
) ...? I believe that in Iceland they yet are found. Old English had a
) bunch of words besides these that were adj's that also behaved as
) prep's; the one left is "near", but on the other hand "du" was pickd
) up: "the honor du me", "tomorrow this is du". It sumtimes seems to me
) that in old Germanic the prep's were an open class, with rules for
) making one from adj's or other words.

"RANDY J. LAPOLLA" ( wrote:

) In the Qiang languages (Tibeto-Burman family, Sino-Tibetan stock)
) there are systems of verb prefixes that refer to geographic landmarks
) such as "towards the river", "toward the mountian", aside from
) "normal" references such as "towards the speaker", etc. Most
) interesting is that these prefixes also mark achievement vs. state,
) and also perfective vs. imperfective.

Lee Hartman ( wrote:

) The following is probably not exactly what you are looking for,
) but it does bear some similarity to your example of
) prepositions that refer to the nearby river.
) Indonesian has four words for north, south, east, and west,
) -- al monomorphemic so far as I know. But northeast and northwest are
) respectively _timur laut_ and _barat laut_,
) literally east sea west sea
) (Southeast is _tenggara_ -- monomorphemic? --
) and I haven't yet found a southwest.)

Thank you all for your very interesting comments.

Kai von Fintel
Dept. of Linguistics & Philosophy
MIT, Cambridge, MA 02139

Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue