LINGUIST List 4.303

Sun 25 Apr 1993

Sum: Indirect Speech

Editor for this issue: <>


  1. , directly vs. indirectly reported speech - summary

Message 1: directly vs. indirectly reported speech - summary

Date: 23 Apr 1993 09:34:05 -0500directly vs. indirectly reported speech - summary
From: <>
Subject: directly vs. indirectly reported speech - summary

My request was for languages which always or usually restrict
reported speech to the directly reported variety, saying
essentially (a) rather than (b):

 (a) He said, I am hungry. (direct)
 (b) He said he was hungry. (indirect)

I was especially interested to see whether anyone would claim
grammatical impossibility of indirectly reported speech for a
language. Respondents were generally reluctant to assert that a
given language absolutely did not employ indirectly reported

Languages which always or usually employ direct speech:

- Navajo and some other Athapaskan languages (some claim indirectly

 reported speech is not possible in Navajo; directly reported
 speech does not have to be exact quote; on Navajo, thesis of Mary
 Ann Willie (Univ. Microfilms) and dissertation by Ellen Schauber
 (Garland) were mentioned)
- Biblical Hebrew (almost always direct)
- Dari (always direct)
- American Sign Language (prefers direct)
- Nahuatl (always direct; has means of disclaiming that exact words
 are being reported)
- Plains Cree and other Algonquian languages (always direct)
- Crow (direct only, in respondent's experience)
- Creek (direct only in natural speech; indirect in elicitation)
- Cubeo and perhaps other Amazonian languages (direct only)
- Sanskrit (usually direct)
- Kiowa (direct only in current corpus of respondent)

Languages which probably should be examined with respect to this

- Japanese

Much thanks to respondents (hope I haven't missed anyone!):

Peter Bakker
David Bergdahl
Andrew Carstairs-McCarthy
Karen Emmorey
David Gil
Eloise Jelinek
Frances Karttunen
Mai Kuha
Jeff Lansing
Jack Martin
Mike Maxwell
Nicholas Ostler
Linda Rashidi
Peggy Speas
Laurel Watkins
Robert Westmoreland

A number of related issues arose:

a. Thought reported as speech. This seems to be common in Biblical
Hebrew. David Bergdahl called my attention to the article on
current Am Eng "like" in_American_Speech_ Fall 91, which suggested
that one function of the BE + like + S construction in narratives
is to signal that the complement reports thought or approximate
speech ("So I'm like, Are you saying I didn't read the

b. The question of whether, when a language employs directly
reported speech grammatically, we may conclude that the speaker
intends to report the exact words of the third party. This is an
especially important question in the case of languages which may
employ only directly reported speech. Here the use of evidentials
and disclaimer formulas like "it is said" are relevant, but it is
not clear in every case whether the use of such a device, in the
case of reported speech, signals that the reported speech act
itself is hearsay, or the words employed in the reported speech

 A narrator may employ intonation shifts (as happens in the case
of Dari; thanks to Linda Rashidi for this very relevant
observation) to signal that the exact words are being reported.
But does that mean that if there is no intonation shift, only
approximate wording is being reported, or that the issue is

c. Whether the presence of a complementizer (such as NT Greek
'hoti' or 'ke' in Dari) has any effect on whether the reported
speech is direct or indirect; apparently in Dari and NT Greek
(Koine) it does not.


 There was a posting--unfortunately my copy was lost, so I
can't credit the author--regarding the possible relevance of Julian
Jaynes controversial theory expounded in _The Origin of
Consciousness in the Breakdown of the Bicameral Mind_ (Houghton
Mifflin 1976) to this issue. The idea was, as well as I can
remember it (I am sure I am not doing it justice), that ancient
languages may have tended toward directly reported speech prior to
the mental/cerebral change in humans posited by Jaynes.

 Jaynes's book is interesting but IMO hard to take seriously,
since uniformitarian assumptions are usually considered important
to the scientific method. In any event, it is not only ancient
languages but also a number of languages spoken today, as this
inquiry has revealed, that prefer directly reported speech.

Again, thanks to all who responded, & apologies to any I may have

Ken Miner
-- | Nobody can explain everything to everybody.
opinions are my own | G. K. Chesterton
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue