In closing, I think it will be helpful to present a condensed review of the criticisms I have made against personality type theory, at least in the form it is taking in psychology today. Although several of this points are related, they constitute distinct and separately identifiable problems for the underpinning presuppositions of personality type inventories. And it is this relative independence of so many serious difficulties that suggests just how far from reliable the psychology presented in such personality theories really is.
-- Naivity in Measurement: Type-trait theories work from concepts of dispositions that generally have broad meanings which have been interpreted different ways in the philosophical discourse of moral psychology, but have often inherit limited, impoverished, or otherwise idiosyncratic construals of these dispositional traits that were fixed earlier in psychoanalytic discourse. Thus, their practitioner's belief that these questions will elicit evidence of just these traits is both subjective and unsupported on its face and often highly distortive. There are several problems here: (A) the questions may be phrased in ways that carry skewed implications about the trait they are being used to measure; (B) a `yes' or `no' answer to the questions could in some cases be evidence not of the disposition which the term actually refers to, but of several others; (C) there may be other questions or observations that would be much better indicators of what this trait-term means; (D) the very concept of the disposition being measured for may be unclear or gerrymandered, the result of a mistaken interpretative focus in earlier moral psychology, and so this single `trait' is actually a partial conglomerate of several other more basic dispositional features.
-- Relevance and Hidden Variable Problems for Factor-Analysis: Type-trait theorists tend to insist that their results are not philosophically criticizable because they are at least potentially supported by rigorously scientific methodology. The methodology in which they place such faith is regression and factor-analysis to determine if correlations are relevant, if trait-continuaa are orthogonally independent of each other, and if the traits they have selected to measure are `real,' or actually non-independent parts of other traits, or linkages of multiple more basic traits, etc. But this method cannot by itself assure us that other hidden variables would not split observed correlations, link factors formerly thought to be independent, or produce more significant correlations. In addition, it cannot assure us that the variables the analysis has identified are relevant for personality, since that concept has a normative dimension that may reduce the importance of factors the models include, or point to the importance of factors it has not even tried to test for.
-- Leveling: The type-trait approach treats what may be intrinsically different kinds of factors that help account for consistency of behavior indirectly in interaction with situations, but treats them all as if they were just the same kind of thing of thing --namely traits or types (complexes of traits)-- and thus metaphysically on a par. In particular, these theories tend not to recognize hierarchical differences, i.e. that some `traits' may constitute not tendencies or attitudes simpliciter, but tendencies or attitudes about other traits --a point brought out in recent moral psychology..
-- False Neutrality: Because they mix together fundamentally different kinds of dispositions, type-trait theories inevitably include --right along with traits that describe different cognitive approaches and interaction styles with no apparent ethical implications-- several other trait-contrasts that reflect moral character, or the `virtues and vices' of the person. Yet because contemporary type-trait theories are at pains to proclaim that the results of their analysis are passing no judgment on the individual, they ingenuously construe every difference in disposition as a `gift' and suggest in practice that none is inherently `bad' or less likely to contribute to human flourishing. This ignores basic insights of the virtue-theory tradition in moral psychology, which recognizes the ineliminably evaluate content of a distinct set of dispositions that are related to choice.
-- Omission and Bipartitism: Because historical biases unconsciously derived from inadequate moral psychologies influence the selection of relevant traits and types for testing, and the interpretation of what kinds of questions will elicit which traits, type-trait theories of personality tend to leave out altogether, or at least dramatically underemphasize, certain relevant trait and trait-complexes, in particular those relating to volition in the sense distinct from both affect and detached cognition. Type-trait portraits of the `personality palatte' thus typically imply no room for a `middle part of the soul,' and revolve around a global bipartite polarization of the rational vs. the emotional.
-- Determinism: Because of their historical origins in deterministic moral psychologies, type-trait theories of personality implicitly leave no room for alternate-possibilities freedom or `liberty' in the determination of one's character, if not also in outward action. Like Leibniz, they assume that differences in attitude and behavior between individuals in similar situations must have `sufficient reason,' which can only be an innate cause rather than libertarian choice.
-- Psychic Alchemy: Taken together, this inherent deterministic tendency in type-trait approaches, their reduction of all relevant dispositional attitudes to one generic level, and their faith in factor analysis yield the sense --which runs implicitly throughout this typology discourse-- that what is being discovered in the analysis of personality traits and types is a kind of `period table' of the basic elements of the soul: just as chemical compounds are the result of the combination of basic chemical elements, so different personalities are the result of a combination of basic dispositional elements, and the type label applying to an individual is almost analogous to a chemical formula. Not only, as we have seen, is the `palette' of dispositional `colors' from which the personality is painted in these theories incomplete; the deeper problem is that unlike molecules, personality may not in fact be a combination of elements at all. This whole paradigm, paradigm, with its associated assumption that we can `prime factor' the soul into a set of basic components, may be fundamentally in error.
The model on which typology approaches rest is thus one of personality as a `painting' made with an array of primary `trait-colors,' which determines how it will look (perform) in the `lighting' of different situations. This model derives from the empiricist era of moral psychology, and is thus philosophically controversial and cannot be taken for granted. If unity-of-character accounts such as those often found in virtue moral psychologies are more correct, it may be misleading to such an extent that the error colors all subsequent interpretation of data gathered, and cannot reveal itself empirically. Statistical analysis of answers to subjectively designed questions that may only inadequately measure an arguably incomplete and poorly conceived set of traits can hardly contribute anything interesting towards resolving the underlying philosophical disputes at stake here.
In closing, let me return to the point from which we set out. Some philosophers today believe that philosophers as a group (especially those working in ethics) should recognize and defer to existing `knowledge' discovered in psychology (and the sciences of the mind generally), and therefore limit their theories to those that fit with or are practicable within the frameworks set by psychology and cognitive science. The case of personality type theories, however, shows how backwards this proposed standard is. Philosophers must deal with empirical findings, but neither ethics nor the underlying metaphysics of personhood can be circumscribed in advance by supposedly scientific theories that always embed philosophically controversial assumptions. For sometimes psychologists even erect entire edifices on bits of outdated metaphyics, distorted mutations of once-clear concepts, and threads of flawed moral psychologies, sewn together hodgepodge in a tangled skein that only the philosopher can hope to untangle and follow back out of the labyrinth to their sources.