terry,
you are asking some questions which i hope i can answer, though they may
still be conceptual shorthand as there usually is much more to the answers.

1. What is a "space of possibilities" - what is it as a concept and what is
its epistemological status relative say to a 'thought', a 'written down and
carefully defined concept',  a 'negotiable idea relativistically ideated on
'a person by person basis', etc. I'm not asking about its meaning. I'm
trying to understand what kind of concept it is.

to me, a space of possibilities is a number of expectations of what could
happen, a number of as yet not chosen actions, a number of conceivable
directions for a development, a number of possible artifacts -- organized in
some way, hence the word space (which implies neighborhood relationships and
structure -- again, not of things but of as of now unrealized or open
places).  mental spaces are internalized conceptions that relate to the need
to explicate the choices one is requested by others to explain.  if you read
julian jaynes about the rise of consciousness, you can see that even such
mental spaces are secondary to the spaces we create in social discourse, in
language.  take a blank piece of paper.  before you write something on it,
you can write anything on it -- within the confines of the paper, provided
you have a writing utensil, and since paper is marked socially (is
manufactured, comes categorized, and with an expected use) -- within the
limits of what you can understand and perhaps communicate to others.  (as
such, this has nothing yet to do with representations).

2. Do you see  a difference, conceptually speaking,  between a 'social
phenomenon' and other phenomena?

sure.  a social phenomenon is constituted in the understanding that its
participants have of it and the enactment of this understanding coordinates
these participants' behavior.  an individual phenomenon does not require
others' participation.  for example, when you get up and night and find your
way from the bedroom to the bathroom, this reality construction is entirely
yours.  but a family, money, government cannot be understood without
understanding the understanding of other participants, i.e., family members
conception of family, business partner's conception of money, and voters',
politicians', and officials' conception of government.

3. I find I can have conversations in my head, taking different roles,
arguing something this way and that as if it were real. Others seem to be
able to do the same. I and many other designers seem to do this and similar
'as if' juggling with ideas about possibilities as a normal part of
designing. Your concluding sentence seems to suggest that this isn't proper
in some way. Please can you clarify.

i wouldn't say that this is improper at all.  it is what we all do as part
of the operation of consciousness:  imagining oneself doing things in all
kinds of situations.   but conversations in your head are secondary to
actual participation in conversations with other people.  i have always said
and written about the fact that one important requirement for human-centered
design is the ability of second-order understanding, the understanding of
the understanding of other stakeholders in design, users, for example.
people who cannot do that cannot do anything for or to others unless they
behave and think similar or identical to them.  the latter is built into
rationalism, the belief that everyone needs to think the way everyone else
thinks and if they don't then they are in some ways inferior or defective.
it is also the cause of interpersonal and international struggles, blaming
women to be irrational,  or opponents to be scum or in the extreme
terrorists.

klaus

===snip
i think you overvalue models of emotion and feelings in processes of design.
surely emotions and feelings are essential in human responsiveness to
anything, especially in the evaluation of alternatives.  but what this
assertion misses is the ability to create spaces of possibilities without
variables.  in my analysis these spaces are created largely in language,
which is a social phenomenon, not in a notion of cognition as an
individualistic phenomenon.  in language we have metaphors, generative
vocabularies, and we test them out in conversation and dialogue, not in any
one individual's head.
klaus