Mike, it's my pleasure to enliven discussion! And thanks for posing some
questions that keep my enthusiasm in line.
1) Regarding your first question--posed to Nacho (hola, Nacho!)--let me
my own name to the list of private speech researchers who have failed to
the stream of private speech using "word meaning". While I have taken a
stance toward my fellow researchers for not incorporating conversational
discourse analysis into their speech analyses (a stance most strongly
in the Robbins, et al. chapter), I, too, am guilty of not using "word
meaning" as my
unit of analysis. Until very recently, I had no idea how to do so, or what
analysis might even look like.
Also, I do not assert that the beautiful new book on private speech edited
Winsler, Fernyhough, and Montero (and it is, indeed, beautiful!) is devoid
"interest" in meaning--what is missing from our collective work on private
I submit, is an "applied analysis" of meaning.
2) As for the role of the "utterance" in the analysis of word meaning, the
problem confronting any speech analyst is how to divide the stream of
into units. The units will depend, of course, on one's purpose. There are
occurring breaks--e.g., pauses, silences--that present themselves as "ripe"
segmenting the stream, but analysts disagree as to the length of time that
be used as a standard measure. Thus, some use pauses lasting only 2
and others use pauses lasting 3 seconds. Depending on the speaker's speed
of uttering, this could mean the difference between segmenting the units
single sentences or segmenting them into multiple sentences. And if one
a really stringent criterion, I could see the stream of speech being
into utterance units consisting of nothing more than individual words or
Which linguistic elements was Bakhtin referring to with regard to an
A word? A phrase? A complete sentence? A lengthy monologue lasting several
sentences? This technical issue seems to me to present a basic problem for
any analysis of speech.
Now, if we decide in advance which linguistic units we want to isolate
confines of an "utterance"--let's say a complete sentence--we could do so
adjusting the pause interval to an appropriate length, and by using
structure as a measure, and by using intonation as a measure (I use a
intonation to identify the end of a question-form, "falling" intonation to
the end of a declarative-form, and a "rising-falling" intonation to demark
of an imperative, or command). Such a method puts the analyst in control of
the linguistic units that are then regarded as an "utterance".
Because Bakhtin was interested in the interaction of voices, and therefore
approached the stream of speech from the vantage point of a dialogical
he focused on the utterance as a sociolinguistic unit. Here I see a huge
between his approach and a conversational approach. The biggest stumbling
block to bringing these two approaches together, as I see it, is how one
the relationship between the initiation-response function and the
embodied in an utterance unit. Stubbs, a conversational analyst [see
(1983). Discourse analysis: The sociolinguistic analysis of natural
Chicago: The University of Chicago Press], defines initiation-response as
interactive unit in which an utterance initiated by one person is met with
responding utterance from another person. That is, dialogue is conceived as
a two-utterance entity. However, my reading of Bakhtin (particularly, The
Imagination) leads me to understand that he viewed adult speakers as having
learned to shift conversational voices from *between* utterances (and
*within* utterances (and individuals). That is, he conceived of dialogue as
utterance entity in which one utterance is voiced externally (explicitly)
and the other
is voiced inwardly (implicitly). If that's so, how are these two approaches
For the past year, I have been working with some student assistants who
been helping me to segment conversation into sequences of utterances, and
it turns out to be incredibly difficult to agree on what constitutes a
if you push the analysis hard enough. For example, looking at a play
in which a lone, older child is talking to himself aloud as he fantasizes,
tended to regard the whole play episode as one long conversation. When I
pressed them to divide the whole episode into sub-topics, this could be
but they complained of its arbitrariness because the sub-topics could be
even further into smaller conversational sequences. When I then took it
to push the analysis as far as it would go, I discovered--to my initial
the lowest granular level was the individual utterance, not a two-utterance
Bakhtin seemed to be right, I concluded, for the simple reason that every
utterance both *responds* to the prior utterance and also *initiates* the
That is, it links both backwards and forwards. The only possible exceptions
are the initial utterance in a sequence, and the final one.
So, Bakhtin's notion that every utterance "anticipates" a response from the
makes sense to me--with inner speech, it is possible to simulate the
response before one utters--enabling the speaker to adapt the utterance
I'm afraid that this is where my thinking comes to an end. More work to do
3) Finally, with regard to published work that folks can read, I suggest
my latest chapter in the Winsler, Fernyhough, and Montero volume [(2009).
C. Fernyhough, and I. Montero (Eds.), Private speech, executive
functioning, and the
development of verbal self-regulation. New York: Cambridge University
That's where my most developed thoughts on the topic can be found. I will
a "pdf" of that chapter below, but PLEASE go out and buy that book--it's
Also, as requested, I will supply in a subsequent email a "pdf" of my
chapter in the
Robbins and Stetsenko volume---ANOTHER magnificent book! [(2002). D.
& A. Stetsenko (Eds.), Voices within Vygotsky’s non-classical psychology:
present, future. New York: Nova Science Publishers.]
Thanks to all for your indulgence. I hope I have provided some food for
(See attached file: Feigenbaum_Development of Communicative Competence.pdf)
Best wishes, Peter
Peter Feigenbaum, Ph.D.
Associate Director of Institutional Research
Bronx, NY 10458
Phone: (718) 817-2243
Fax: (718) 817-3203
Sent by: "eXtended Mind, Culture, Activity"
10/31/2009 08:39 Re: [xmca] The Ubiquity of
PM Unicorns: conversation
Please respond to
m; Please respond
As usual, tardy to the party. Very interesting comments on private speech,
thanks for starting it Peter.
A couple of questions for different xmcaonaughts:
What about Peter's observation, Nacho?
"Personally, I don't know of one private speech researcher who uses LSV's
concept of word meaning when analyzing speech data--and to me, that
something is very wrong." That beautiful new book you have published on
private speech is devoid of an interest in meaning? Really?
And Peter, what about Andy's comment, which also struck me, that when you
talk about your data, your unit of analysis appears to be the utterance, a
la Bakhtin. And Bakhtin makes central to his analysis the idea that words
are always characterized by addressivity. Seems like grist for your mill.
But utterance is absent from your jpeg figure, and Bakhtin from your
discussion thus far. And does conversation start with the first
word spoken by the child? All Samoan children are said by their parents to
have as their first word, shit (or so ethnographers have claimed). What
would that imply for addressivity of first words, I wonder.
Someone asked about availability of published work for folks to read; might
you post your article from the Robbins et al book? And, of course, we would
be glad to post a video session and transcript. The topic seems of great
importance, and not unrelated to the once and future thread of the
blocks" methodology. Any private speech observed there, Paula?
Thanks again for the enlivening discussion.
On Fri, Oct 30, 2009 at 10:08 AM, Peter Feigenbaum
On Oct 29, 2009, at 5:03 PM, Andy Blunden wrote:
I see you use utterance as a unit of analysis. In the previous mail
you referred to using "conversation", but utterance is surely the
unit of which conversation is made up. Do you have Bakhtin's idea in
mind for "utterance" at all? Otherwise the notion of "turn taking"
in private speech is very challenging, isn't it?
All sounds fascinating,
xmca mailing list