136
John B. Haviland
or
‘rabbit essence,’
etc.
–
underscores the profound referential
indeterminacy
of linguistic behavior. Perhaps more to the point is Zgusta’s analogy
(Zgusta 1971: 25–26) with trying to discover the meanings of traffic signs
(in a system like the European one), but only on the basis of observing the
regularities in drivers’ behavior. Perhaps, speculates Zgusta, one could in
time decipher the meanings of, say, the red, yellow, and green signals of a
traffic light by direct observation; but the meaning of a “great capital H on
a rectangular shield (which means in many countries that there is a hospital
not far away)” would be much harder to divine, since such signs stand in
many different kinds of locations and “a uniform effect on the behaviour of
other drivers is hardly observable.”
Here is a less fanciful example from the annals of real field lexicography.
In 1770, Lt. James Cook and his crew collected wordlists from the Guugu
Yimithirr language, spoken near what is now called Cooktown, in north-
eastern Australia. (One word was gangurru, the name for a particular spe-
cies of what the world now calls kangaroos). Collating the shared entries of
different observers, one can see precisely that referential indeterminacy of
the gavagai variety plagued these early lexicographers. Thus, under the
gloss ‘branch (with buds or stalk)’, the ship’s illustrator Parkinson has
maiye, Banks the botanist writes maye butai (adding the annotation ‘with
leaves’) or mayi bambier. Based on the modern language, I assume that
these expressions are based on the word mayi ‘edible plant’ – so not just
any old branch is involved – and more specifically mayi bambiir ‘the (edi-
ble) fruit of the mangrove species called bambiir’. The other “name” Banks
records is plainly the expression mayi buday which is really an entire sen-
tence that means “the edible part has been eaten” or “someone ate the
fruit.”
4
Cook’s journal entry shows he was painfully aware of such Quinean
problems of lexical elicitation.
…the list of words I have given could be got by no other manner than by
signs enquiring of them what in their Language signified such a thing, a
method obnoxious to many mistakes: for instance a man holds in his hand a
stone and asks the name of [it]: the Indian may return him for answer either
the real name of a stone, one of the properties of it as hardness, roughness,
smoothness &c, one of its uses or the name peculiar to some particular spe-
cies of stone, which name the enquirer immediately sets down as that of a
stone.
(Cook’s journal, see Cook 1955)
Chapter 6 – Documenting lexical knowledge
137
Part of the problem, clearly, is in a primitive model of both reference and
ostension: what you can pick out by pointing, or what you can show “the
Indian.”
A very different model of “exemplification” is advocated by J. L. Austin
in “A plea for excuses” (Austin 1961). Faced with a pair of expressions
(famously, in Austin’s case, the apparently similar by mistake vs. by acci-
dent) one elucidates the difference in their meanings by constructing a careful
example of when you would use the first expression but not the second, and
vice versa. In such a method one points not at things but at contexts of use.
Contexts themselves can be crucial in accessing lexical knowledge. In
trying to recover words from the native Barrow Point language of the late
Roger Hart, he and I worked largely through Guugu Yimithirr, a second
language for both of us (see Haviland 1998). We would often search –
sometimes quite naively – for the Barrow Point equivalent of a Guugu
Yimithirr word. Even looking for the names of plant or animal species,
however, we were often stymied, partly because the flora and fauna of Bar-
row Point were frequently different from those of Cape Bedford, more than
a hundred kilometers to the south, but partly because the environment in
general was just wrong. Roger had learned his tribal language before he
was removed from his family around the age of six. I first heard him speak
the language without hesitation, however, sixty years later. After a long
trek back overland, he and I stumbled out onto the beach where he had been
born. The country he had not seen for sixty years, its trees, rocks, and ani-
mals, seemed to speak to him in his childhood tongue, and he was only
there able to respond fluently.
Reference – or more precisely those aspects of linguistic expressions
that render them useful for achieving reference – though the staple of most
modern formal semantics, is of course an inadequate basis for understand-
ing meaning in an ordinary sense. The traditional notion of “connotation,”
for example, is based on the intuition that different words can in some
sense “refer to the same thing” without, thereby, “having the same mean-
ing.” This is not the same as Frege’s classic distinction between sense
(what an expression means) and reference (what it just happens to refer to,
as a function of what it means) where two different expressions, with dif-
ferent senses, can happen to refer to the same individual. Zgusta’s some-
what quaint example is the lexical triad ‘decease’, ‘die’, ‘peg out’ (the last
in my own dialect of English would be something like ‘check out’ or per-
haps ‘go belly up’). Zgusta (1971: 39–40) cites Armenian as a language
which has exact counterparts (va xanvel, mernel, satkel) for these English