Tolowa dative/benefactive

[Tom Givon]

In languages where the verbal meanings 'say' and 'tell' share the same stem
(without derivational affixes), one should not call the stem itself 'the verb'.
Rather, one must take each sense of that stem and test it for its syntactic
properties--in its distinct syntactic frame. This is true in Ute, Tolowa, and
many other languages. In such languages, most commonly such a verb- can fit
into four distinct syntactic frames:

   (i) With a non-equi clausal complement:
                       'She said that Matt had left'
   (ii) With a non-equi clausal complement and a DAT-human nominal object:
                       'She told/said (to) Sarah that Matt had left'
   (iii)  With an equi clausal complement and a DAT-human nominal object:
                       'She told Matt to leave'
   (iv)  With an ACC and a DAT nominal objects:
                       'She told Matt a story' ('She told the story to Matt')

The syntactic question one could investigate then, whether the DAT argument is
coded as DO, applies only to classes (ii), (iii) and iv). In general, all these
three tend to go by the same rule, but that has to be tested language by
language.   TG

========================


dryer at buffalo.edu wrote:

> I think one needs to be careful about using a verb meaning 'tell' to see
> how a language treats "dative" arguments, because in many languages the
> verb meaning 'say' is morphologically and syntactically intransitive and
> the verb meaning 'tell' is morphologically and syntactically
> monotransitive, not ditransitive.  In Kutenai, for example, there are a
> variety of tests that show that the verb for 'say' is intransitive, so that
> the thing said, while semantically an argument of the verb, is
> grammatically an oblique (e.g. the verb for 'say' takes nonspecific subject
> marking, something that is only possible with intransitive verbs; and to
> relativize 'what is said', one uses a relativizing construction that is
> used for relativizing obliques, not the construction for relativizing
> subjects and objects).  And one forms the verb for 'tell' by adding a
> transitivizing suffix to the verb for 'say', so that the verb for 'tell' is
> simply a monotransitive verb, and thus different from the verbs meaning
> 'give' or 'show'.
>
> On the other hand, if Tom is right that the Tolowa verbs for 'show' and
> 'teach' are like the verb for 'tell' (especially when they take nonclausal
> complements, as opposed to the meanings 'show that' or 'teach that'), then
> these verbs would support his claim.
>
> Matthew Dryer