20
thought (established or not). In relation to reasoning, it is widely recognized how easy
it is for someone to reason to different conclusions from those of her fellow if their
two foundational belief systems are different.
But what meanings count as immediate? Could the set of human primitives be
specified? Suppose we are asked to classify the words set, electron, water and pain in
either one of the following two mutually exclusive sets: (i) the set of primitives P; (ii)
its complementary P'. It seems that set falls in P electron in P'. What about water and
pain? It might seen obvious that H2O being analyzable to its constituents parts and
these in their turn to electrons (among other similar types of entities) should be placed
in P'. Still, I can think of no one who would not place water in P. Moreover, set and
water are two quite different types of primitives. Set can be a primitive for a
mathematician whereas water is a primitive to anyone. What about pain? It might
seems obvious to everybody that pain should be put in P, but Dennett (1978) has
given a nice, although sketchy, model of pain; it follows that one should put pain in
P'. So, we observe that three out of the four concepts can be placed either in P or in P'
depending on whom we asked to classify them.36 So, primitives are of two types
linguistic and pre-linguistic or sense ones. The latter are inseparable from the
corresponding organism concerned. The former depend on the particular individual H
possessing them. To be noted that as soon as linguistic primitives are expressed they
become data and they may only acquire meaning if processed by another human. In
summary, ΠH is both subjectively and community dependent. More generally, one
can convincingly argue that ΠE is both individual and species-specific. It becomes
obvious now that a ΠH or a ΠE constitutes an extremely complex system. To get a
better idea of this complexity consider the following further characteristics and
examples.
First, the fact that the meaning of the elements of a ΠH is immediate does not
imply that they are necessarily atomic (i.e., not able to be split into simpler ones).
Qualia constitute an important class of atomic ΠH. The second, related, feature is that
ΠH can change with time. A human may split some of the elements of a ΠH into
simpler ones or she may actually abandon some. An important special case of ΠH
modification is usually involved in conceptual change (e.g., Vosniadou & Verschaffel
2004). Third, TonE allows acquired primitives in contrast to Fodor’s (1994) strongly
innatist account according to which primitives are inherited as part of the structure of
the brain. Fourth, the systems of ΠH are not in accordance with either Fodor’s (1983)
modularity thesis or Tooby and Cosmides (1992) massive modularity one. Contra
Fodor systems of ΠH are interrelated via thinking or understanding and therefore not
encapsulated. Contra massive modularity, a large number of ΠH systems are
ontogenetically specific and therefore not adaptations. Fifth, a ΠH system is
fundamentally different from formal systems of primitives like those identified with
pixels or curves (see Schyns et al. 1998 for a discussion). Sixth, since primitives are
meanings, a ΠH is a subsystem of an Nm. Moreover their characteristics imply that
they are quite diverse and diffuse systems of Nm . This claim is in accordance with
fMRI results of brain’s semantic organisation (Bookheimer 2002). Seventh,
primitives are age- individual- cultural- and species-specific. An example of age-
related primitives is ‘electricity’; it can be a primitive for a toddler but not a primitive
for an electrician or quantum physicist. With respect to individuals, 'water' was a
primitive for my grandmother throughout her life but it is not a primitive for those
knowing that 'water' can be thought of in terms of H2O. Actually, what may be a
primitive for someone may be part of a very complicated theory for another. As an