(cf. Figure 2). Identification of the multiple, distinct but related senses of such
phenomena is a requisite first step. The explicit use of the time variable in defining
them is a useful reminder of their evolutionary characteristics and a necessary tool in
the gradual transition from our vague conceptual apparatus to a less imprecise theory.
Evaluation criteria -by revealing weaknesses- constitute a significant tool in the
development of a scientific field. Newell’s final list of thirteen evaluation criteria
(constraints in his words) fall into two categories: performance requirements and
construction constraints (Newell 1990, pp. 18-21).
A revised version of those criteria was used by Anderson & Lebiere (2003) as
the Newell test for the evaluation of cognitive theories.4 Although there was
agreement of both the target article authors and commentators that the N-test is not
complete (e.g., Agassi 2003, Anderson & Lebiere 2003, Gelepithis 2003, Sirois 2003,
Taatgen 2003), and some argued it is not fully appropriate (e.g., Young’s (2003)
proposal to substitute compliancy for universality, Wang et al.’s (2003) questioning
of its theoretical attainability), no-one made the point that the N-test is actually more
inadequate than the original list as a basis for evaluating a unified ToM.
The argument is simple. Newell’s original list includes two criteria that are
absent from the proposed Newell test and which Newell himself considered
necessary, namely, “operate autonomously, but within a social community” and “be
constructible by an embryological growth process.” (e.g., Newell 1990, p. 20). In
addition, the first criterion on the original list does address a major human capability
while reflecting a performance requirement (ibid. p. 20). In contrast the first criterion
on the proposed N-test (computational universality) provides straightforward grading
but leads to contradictions (Young 2003).5 The inadequacy of the original list itself is
due to the omission of field requirements like invariant laws and minimum
vocabulary. Simon (1990) has elevated the discovery of invariant laws to the status of
the fundamental goal of science and has suggested two laws of qualitative structure
and four quantitative findings as invariants. Surprisingly, although Newell (1990)
strongly believed that the computer hierarchy is an invariant law, he did not included
invariant laws as a criterion of a unified theory of mind. I fully agree with Simon. In
section 2.1.2.3 I propose the communication understanding principle (CUP) as a
psychological invariance.
A minimum vocabulary is a clear sign of the maturity of a science (Russell
1948). Even if not completely attainable, their quest is useful in accordance with the
principle of parsimony. Cognitive science needs a minimum vocabulary to serve as
its descriptive base and substantially decrease the bewildering multiplicity of terms
used. In section 2.3, I propose a set of words that seem to constitute a minimum
vocabulary for a small but significant part of cognitive science. It is a consequence of
the theory outlined in section 2. Concluding on the evaluation criteria of a UTM, I
propose that invariant laws and minimum vocabularies should be an integral part of
such a system.
The next section proposes an account of mind that is both in accordance with the
constraints identified and breaks away with the long tradition of the intrinsic
individuality of mental phenomena traced back to James’s conception of Psychology
as the study of “finite individual minds”. In addition, although there are many
specific, contemporary research programs dealing with ‘thinking’, ‘representation’,
and ‘communication’, there was none that had attempted to synthesize all these areas
as well as a considerable part of the disparate underlying research and propose a,
much needed, minimum vocabulary for this cluster of core mental phenomena. This is
the objective that next section approximates. Specifically, in the introduction to