[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [xmca] Types of Generalization: concepts and pseudoconcepts

Are you the Peg Griffin who was here at ucsd?
On Sep 14, 2009, at 5:36 PM, Peg Griffin wrote:

It's just a part of ecological validity (EV), Andy.

Who's to say that the "p" "not p" version of the Wason task is the
genetically primary example of it?
Maybe the retail store task is closer to that.
And maybe the mathematizing of it was a social move for purposes unrelated to the properties of the task -- maybe to advance the intellectualization of management, maybe to fortify mathematics, maybe just to build an SES sorting
machine far away from retail clerks' interests.

Here's a little bit on EV:
When you work with a finding from the experimental setting or the school
setting, you know something but there may be more to it.
In general, in non-EV work, you can interpret the "correct" with some degree of comfort but it is hard to be sure what is going on with the "incorrect."
In non-EV settings, we are unlikely to get at the motives and goals
organizing the subjects -- the whole task -- since the division of labor
with the tester can fudge this beyond observation, recognition, and
And, what you see in the one may not be the "whole" you need or the
interpretation you build may not be the only possible one or the better one. When your training is non-EV (or vice-versa), transfer or generalization is
not as simple as folks once thought.

So EV research is about the following sorts of explorations:
What can you learn about a theory or an aspect of a theory if you find it
"in the wild."
How are its properties and processes different in more naturally occurring
Do the results of the search and the outcome of more "ecologically valid" experiments cast doubt on or fortify the theory or different versions of the

So, the post I referenced suggested to me that the blocks experiment might have an "in the wild" version when the leading activity is trade or capital consumerism. (Sakharov's telling the subjects the task is about names of toys of children far away -- that could be a way to try for a play leading
activity ...?)

What happens if analogue experiments to the blocks one are built and they
include work (as trade or capital consumerism) as a leading activity?
Will it cast new light on the issues about concepts and how they are grown
and how they differ and how they are used?

Suppose people get "trained up" on the trade/consumerism instance then you
give them "transfer triggers" and then give them the blocks.
How do they do on the blocks after this "priming?"

Do the strategies differ for blocks solvers and consumer solvers?

Do the successful folks in one version differ from those successful in
another version based on individual, group, development, or socio- cultural
aspects of their life experience?

Then, do any of the outcomes help with how you theorize, how you interpret others' theories, and how you practically work for success for more people
in a more just world?

I think it was the man I referenced, D'Andrade, who likened his work to doing geology in the midst of a rock slide. (Mike will know who originated
this analogy.)
Ecological validity is a little niche to jump into and keep our heads about
us for a bit.

-----Original Message-----
From: xmca-bounces@weber.ucsd.edu [mailto:xmca- bounces@weber.ucsd.edu]
On Behalf Of Andy Blunden
Sent: Monday, September 14, 2009 4:50 AM
To: eXtended Mind, Culture, Activity
Subject: Re: [xmca] Types of Generalization: concepts and

Fascinating area of research, Peg. I had to use Google books
to figure out what the hell you were talking about!

My only reaction is this: the blocks experiment was
deliberately designed to isolate the tasks from any
realistic references which could tap into life experience.
That was necessary for the "testing" rationale of the
experiment. But by isolating the sensuous content of the
experiment from any meaningful life experience, they make it
difficult for any conceptual thought to enter into the
process. You really have to be a mathematician to get
anything out of it.

But I'm still not quite sure what you were driving at, Peg.


Peg Griffin wrote:
Steve or Andy wrote: " the principle is actually very familiar, for
to modern consumers when they compare similar commodities of
brands and models for desired (and undesired) features, prices, etc."

So, the D'Andrade type experiment is just waiting to be done!  (In
D'Andrade's variation on Wason’s selection task, the scenario
involves a
task for a clerk at a department store deciding about which checks
needed to
be turned over to check if the supervisor had signed off on it.)

Anyone do it yet?

-----Original Message-----
From: xmca-bounces@weber.ucsd.edu [mailto:xmca-
On Behalf Of Steve Gabosch
Sent: Saturday, September 12, 2009 11:05 PM
To: eXtended Mind, Culture, Activity
Subject: Re: [xmca] Types of Generalization: concepts and

Andy, thanks for your response to Davydov on concept formation and
efforts to get us to read Davydov, Vygotsky, Sakharov, etc.  It has
certainly been effective in my case.  And Jay, your comments have
been very helpful.

Two questions on your essay, Andy.

One, what do you mean by "an absolutely non-empirical social factor"
when you say: "The transition from complex to concept is a
and complex process, but one which necessarily involves a complex
leap, in which absolutely non-empirical, social factors enter into
formation and enrichment of the concept."

Two, I am having difficulty understanding how Sakharov block
for bik, cev, lag and mur are not "true concepts" in the way
used the term.  A taxonomy formed out of formal rules can be a true
concept, yes? The Sakharov block test is really just a puzzle where
you have to figure out the taxonomic classification system by
observing the visible attributes of the blocks and figuring out the
only one that can be put into four logical groups.  Yes, the
words are arbitrary and only have meaning to test participants - but
that is the case for any game.  In chess, for example, rooks and
are "concepts" - yes?  If a rook is a concept, then why not bik,
mur and lag?


Here are some details on the Sakharov test and its solution that
help visualize this question of whether the solution groups to the
test are themselves "true concepts".  In discussing details to the
solution to the test the way I do, I am arguing that the solution
groups are "true concepts."  I am willing to be corrected on this,
course.  Perhaps there is a better way to interpret these details.

The 22 Sakharov blocks were very cleverly designed.  No two blocks
exactly alike.  They are comprised of 6 different colors, 5
geometric shapes, 2 different heights (tall and flat) and 2
sizes (large and small). There would be 120 (6*5*2*2=120) different
blocks altogether if a full set of blocks were created out of these
parameters.  The 22 that were selected have the interesting
characteristic of having one and one only possible rule-based
to the challenge of sorting them into 4 logical groups based on
physical attributes.

Since there are 4 groups that these 22 blocks are going to fall in,
one's first impulse is to look for a single parameter that all
share that has 4 variations.  As it turns out, this is impossible.
There is no 4*1=4 solution.  That took some serious design
forethought.  There are not even any clever, obscure alternative
solutions along these lines.

In one of Paula's earlier papers, she reports on a child who, after
deciding that neither color nor shape would work as solutions, began
counting numbers of **sides** of the blocks to see if **that**
parameter fell into 4 groups. It doesn't - they fall into 5 groups.
That little inspiration got me me to try to come up with some other
way of grouping the blocks into 4 logical groups by seeking unusual
parameters, such as numbers of angles, numbers of two-surface
intersections, numbers of three-surface intersections.  However, no
single parameter I have come up with has has only 4 variations. (As
an aside, most of these parameters just mentioned, interestingly,
5 variations - the reason being that all the 6 different geometric
shapes have different totals of these unusual parameters except the
square and trapezoid, which have the same number of these - so
consequently, the total of 5 keeps reappearing).

I don't think it is a coincidence that there are no alternative
solutions. I am guessing that Sakharov very deliberately designed
these blocks to avoid that distraction. This is part of this test's
very clever design.

What makes this test a puzzle even to most adults is that the
requires not finding one parameter with 4 variations, but combining
**2** parameters that each have **2** variations.  I think Paula
this a dichotomous solution (Paula, do I have the right word?).
Running into this principle in the way this test presents it is not
everyday occurrence, but the principle is actually very familiar,
example, to modern consumers when they compare similar commodities
different brands and models for desired (and undesired) features,
prices, etc.  Once one understands this general principle
the parameter variations to figure out the total possible
combinations) and that this is the way this Sakharov-block puzzle
works, the solution becomes completely obvious by just observing the
parameters and counting their variations.  Since the solution seeks
groups, and since there are no 4*1=4 solutions, the one and only
possible other solution would be to find a 2*2=4 way of assembling
groups together.  And wallah!  There the solution is, plain as day
once you see it - tall/flat and large/small.

In theory, if one understands this principle clearly, one could
determine the different groups just by looking at the 22 blocks,
counting and calculating the parameters and their variations by eye,
and do so without picking up a single block.  However, since the
nonsense words are arbitrarily assigned, it would still be necessary
to pick up a block in each of 3 different groups to determine the
precise names that correspond to each group.  There probably are
people who could figure this all out just by staring at these blocks
and arriving at this reasoning, but they would have to be a pretty
experienced puzzle solver to do that in one shot, I would think.
However, there are many very bright people associated with this list
anyone solved or seen the test solved in "one shot," so to speak?
interesting question to ask is, about those that do solve the test -
which solve it **conceptually**, and which stumble on the solution
just a pseudoconcept?)

The question Mike and Paula discussed, and I think David raised,
what procedure or methodology does the test-giver use to guide the
test-taker during the test, is especially interesting.  Which block
they overturn under what circumstances to show the test-taker the
error of their ways during the test, and what other "hints" and
"prods" to they provide as the test proceeds?   (The younger the
child, the more creative prods are needed, from what Paula's
reports!)  This question is interesting on two levels.  One,
obviously, relates to how these prompts influence what the test-
understands and does.  But here is another level to look at this
**what concepts** are guiding the **test-giver** when they are
their prompts?  (And if they are not being guided by "true
then what are they being guided by?)

My point in going into all this detail is to suggest that this
parameter-counting principle is a concept, (or combination of
concepts), and that the solution groups, which themselves are
organized according to this principle, being completely derivative
this overall concept, are necessarily concepts as well.
I am suggesting that these are "concepts" within this
designed system in the same sense that the numbers 1, 2 and 3 are
"concepts" within the number system.

Bik, cev, lag and mur, according to this reasoning, are the made-up
names for specific concepts and are arbitrarily assigned - as are,
ultimately, all words for the things they correspond to.  In this
game, these four nonsense words correspond to the concepts flat-
flat-small, tall-large, and tall-small, which are meaningful
within the game's rules.  These conceptual groups are an integral
of that puzzle's internal taxonomy and its overall conceptual system
even though this puzzle, in many ways, is just about as artificial,
rule-based, experimental, arbitrary and trivial as you could
invent and still get children and adults to make sense out of.  But
lots of cool puzzles are kinda like that.  And this Vygotsky-
concept formation test really is a cool puzzle.

Well, that's my argument for calling these nonsense words "true
concepts" in the Vygotskyan (not necessarily the Davydovian) sense.

- Steve

On Sep 11, 2009, at 1:14 PM, Jay Lemke wrote:

A small follow-up, having now read at least Andy's comments on
Davydov, if not the Davydov itself.

I would agree very broadly with what Andy says, and highlight one
point and note one that is perhaps underemphasized.

Maybe it's because of Davydov's view,  but it seems clear to me
LSV emphasizes very strongly and consistently the key role of
language, and so we ought really want to know more about exactly
the ways in which children and early adolescents use verbal
languages changes as they come to mediate their activity more along
the lines we might call acting-with-true-concepts.

What struck me as very important, that Andy emphasizes (and Davydov
also?) is that the development of true concepts depends on their
in social institutions. This limits the relevance of artificial-
concept experimental studies in ways that would not be apparent in
more purely cognitive science paradigm (or old fashioned empirical-
concept ideology), because the similarity to natural true concepts
is only logical-formal, and not also social-institutional. A lot of
my own students tend to get this wrong, because they identify the
social with the interpersonal, such that there is still a
(in the micro-social milieu of the experiment itself as a social
activity). But not at the macro-social institutional level.

And here perhaps is also a clue to my query about how the modes of
mediation differ across the historical cases (Foucault), the cross-
cultural cases (Levi-Straus), the post-modern cases (Wittgenstein,
Latour), and even the everyday true concept vs. formal scientific-
mathematical true concept cases. The difference arises in and from
the institutional differences. Could we perhaps combine LSV's
insights into how this works in the developmental case (changes in
the social positioning of the child/adolescent), L-S on the
functioning of mytho-symbolic mediated activiity in rituals and
social structuration processes, F on changes in the historical
institutions (medieval-early modern), and L on heterogeneity of
mediation in relation to heterogeneity of actant networks? to
understand better how this institutional context and its processes
play out?

I left out Wittgenstein, but he may help with an intermediate
not the large social institutions, but the game-like activities of
which they are composed.

I'll be looking at Davydov to see what he offers in these terms.


Jay Lemke
Professor (Adjunct)
Educational Studies
University of Michigan
Ann Arbor, MI 48109

On Sep 11, 2009, at 5:51 AM, Andy Blunden wrote:

I have prepared a response to Davydov's book, but it is 4,000
words, so I have attached it in a Word document. But here is a

Davydov claims that in his analysis of the Sakharov experiments,
Vygotsky fails to demonstrate any real distinction between a true
concept and an abstract general notion (what is usually and
mistakenly taken for a concept in non-Marxist thought).

I claim that he has a point, but Vygotsky is guilty only of some
unclarity and inconsistency in his language, and makes the
distinction very clear. And Davydov should pay more attention to
what Vygotsky says about the relationship.

Davydov works with a mistaken contrast between scientific concepts
and the general notions derived from everyday life. Scientific
concepts are by no means the only type of true concepts and
everyday life is full of concepts.

Nonetheless, Davydov has a point. It is evident that Sakharov, the
author of the orignal, oft-cited report evidently is guilty
as charged by Davydov. And no-one seems to have noticed!

Although Paula and Carol are consistent and correct in everything
they say in their paper, they err on one occasion only when they
cite Kozulin citing Hanfmann. It is as if people equate logical
of generalized empirical notions with conceptual thought, never in their own words, but only by means of citing someone else's words.

I think this is the legacy of a lack of clarity in Vygotsky's

4,000 words attached. And apologies for not entering the
of Paula and Carol's paper earlier, but I was not clear in my own
mind on these problems, and Davydov helped me get clear. Better
late than never!


Andy Blunden (Erythrós Press and Media) Orders:
xmca mailing list
xmca mailing list
xmca mailing list

xmca mailing list

Andy Blunden http://www.erythrospress.com/
Classics in Activity Theory: Hegel, Leontyev, Meshcheryakov,
Ilyenkov $20 ea

xmca mailing list

xmca mailing list

xmca mailing list