Using normative correctness as a diagnostic tool reduces the outcome of complex cognitive functions to a binary classification (normative or non-normative). It also focuses attention on outcomes, rather than processes, impeding the development of good cognitive theories. Given that both normative and non-normative responses may be produced by the same process, normativity is a poor indicator of underlying processes.

Humans are not the only animals to deceive, though we might be the only ones that lie. The arms race von Hippel & Trivers (VH&T) propose may have only started during hominin evolution. VH&T offer a powerful theory, and I suggest it can be expanded to explain why there are systematic biases in human foresight.

According to Carey (2009), humans construct new concepts by abstracting structural relations among sets of partly unspecified symbols, and then analogically mapping those symbol structures onto the target domain. Using the development of integer concepts as an example, I give reasons to doubt this account and to consider other ways in which language and symbol learning foster conceptual development.

Carey focuses her attention on a mechanism of concept learning called “Quinian bootstrapping.” I argue that this form of bootstrapping is not dependent upon language or other public representations, and outline a place for language in concept learning generally. Language, perception, and causal reasoning are all sources of evidence that can guide learners toward discovering new and potentially useful categories.

I defend the claim that in psychological theories concerned with theoretical or practical rationality there is a constitutive relation between normative and descriptive theories: Normative theories provide idealized descriptive accounts of rational agents. However, we need to resist the temptation to collapse descriptive theories with any specific normative theory. I show how a partial separation is possible.

Mathematical developments in probabilistic inference have led to optimism over the prospects for Bayesian models of cognition. Our target article calls for better differentiation of these technical developments from theoretical contributions. It distinguishes between Bayesian Fundamentalism, which is theoretically limited because of its neglect of psychological mechanism, and Bayesian Enlightenment, which integrates rational and mechanistic considerations and is thus better positioned to advance psychological theory. The commentaries almost uniformly agree that mechanistic grounding is critical to the success of the Bayesian program. Some commentaries raise additional challenges, which we address here. Other commentaries claim that all Bayesian models are mechanistically grounded, while at the same time holding that they should be evaluated only on a computational level. We argue this contradictory stance makes it difficult to evaluate a model's scientific contribution, and that the psychological commitments of Bayesian models need to be made more explicit.

Patients with physical symptoms for which no organic cause can be found are distributed along a continuum of disease simulation that ranges from a sincere belief of having a serious disease to intentional presentation of false symptoms. The evolutionary hypothesis that self-deception improves the deception of others can explain such a combination of unconscious and intentional production of physical symptoms.

Von Hippel & Trivers (VH&T) argue that people become effective liars through self-deception. It can be said, however, that people who believe their own stories are not lying. VH&T also argue that people are quite good lie detectors, but they provide no evidence for this, and the available literature contradicts their claim. Their reasons to negate this evidence are unconvincing.

I make two points in this commentary on Carey (2009). First, it may be too soon to conclude that core cognition is innate. Recent advances in computational cognitive science and developmental psychology suggest possible mechanisms for developing inductive biases. Second, there is another possible answer to Fodor's challenge – if concepts are merely mental tokens, then cognitive scientists should spend their time on developing a theory of belief fixation instead.

Elqayam & Evans (E&E) argue that we should avoid evaluative normativity in our psychological theorizing. But there are two crucial issues lacking clarity in their presentation of evaluative normativity. One of them can be resolved through disambiguation, but the other points to a deeper problem: Evaluative normativity is too tightly-woven in our theorizing to be easily disentangled and discarded.

Our target article identified normativism as the view that rationality should be evaluated against unconditional normative standards. We believe this to be entrenched in the psychological study of reasoning and decision making and argued that it is damaging to this empirical area of study, calling instead for a descriptivist psychology of reasoning and decision making. The views of 29 commentators (from philosophy and cognitive science as well as psychology) were mixed, including some staunch defences of normativism, but also a number that were broadly supportive of our position, although critical of various details. In particular, many defended a position that we call “soft normativism,” which sees a role for normative evaluation within boundaries alongside more descriptive research goals. In this response, we clarify our use of the term “instrumental rationality” and add discussion of “epistemic rationality,” defining both as descriptive and non-normative concepts. We consider the debate with reference to dual-process theory, the “new paradigm” psychology of reasoning, and empirical research strategy in these fields. We also discuss cognitive variation by age, intelligence, and culture, and the issue of relative versus absolute definitions of norms. In conclusion, we hope at least to have raised consciousness about the important boundaries between norm and description in the psychology of thinking.

The commentators raised issues relevant to all three important theses of The Origin of Concepts (henceforth TOOC). Some questioned the very existence of innate representational primitives, and others questioned my claims about their richness and whether they should be thought of as concepts. Some questioned the existence of conceptual discontinuity in the course of knowledge acquisition and others argued that discontinuity is much more common than was portrayed in TOOC. Some raised issues with my characterization of Quinian bootstrapping, and others questioned the dual factor theory of concepts motivated by my picture of conceptual development.

Commentators raised 10 major questions with regard to self-deception: Are dual representations necessary? Does self-deception serve intrapersonal goals? What forces shape self-deception? Are there cultural differences in self-deception? What is the self? Does self-deception have costs? How well do people detect deception? Are self-deceivers lying? Do cognitive processes account for seemingly motivational ones? And how is mental illness tied up with self-deception? We address these questions and conclude that none of them compel major modifications to our theory of self-deception, although many commentators provided helpful suggestions and observations.