**3. Cognitive models of mindreading-metacognition relation**

Contemporary scientific discussions on the relation between metacognition and mindreading result in several theoretical accounts allowing various combinations and configurations of both mechanisms in the formation processes of beliefs, intentions, and decisions with respect to oneself or others [4]. Cognitive architecture that considers possible configurations of these sets of mechanisms must embrace different functions, knowledge structures, as well as mechanisms that provide access to the information [6]. Yet, it seems that the discussion on the adequacy of each theoretical account is still open, and there is no unambiguous evidence pointing out clear superiority of the specific theoretical account on the relation between metacognition and social knowledge.

For example, Nichols and Stich [28] in their theoretical description of how we access and utilize self-knowledge and other-knowledge propose a hybrid architecture composed of metacognition and mindreading as distinct and innate mechanisms. According to this view, the basis of self-knowledge is formed by two metacognitive mechanisms of self-monitoring nature: one responsible for recognizing and providing knowledge about internal states (our own propositional attitudes) and one for recognizing and providing information on our experiential states [4]. Whereas the mindreading faculty constitutes independent mechanism that deals with the attribution of mental states to understand other people. Since metacognitive and mindreading systems are modular, there is a superior coordinating mechanism that manages the interaction between components [28].

In other theoretical proposals, cognitive architecture on social cognition takes on a more radical form, as the cognitive system is greatly simplified to only one capacity, which is, namely, mindreading or metacognition. For example, Carruthers [4] in his theoretical approach denies that there is an introspective access to propositional attitudes and postulates the existence of only one mindreading mechanism that underlie our social cognition. This view claims that metacognition is only grounded on mindreading and therefore attribution processes of mental states to oneself and other people are results of prior unconscious interpretation. Adoption of such a one-system architecture indicates that knowledge about our own state results only from mindreading mechanisms that accesses a variety of information sources (percepts): incoming perceptual states or quasi-perceptual states [4]. This account clearly implicates that introspection of propositional states is replaced by interpretation without conscious access [4].

Indeed, the idea of one-system architecture ignores conscious access to the information that is, in fact, in contradiction with commonsense observations of the role of consciousness in regulating our behavior. In fact, although our cognitive system in various situations is dominated by unconscious events and shows some limitations, we can consciously correct our theories and judgments to interpret and effectively predict behavior of other people (see, for instance, [29]). In fact, it is important to emphasize that the one-system architecture by Carruthers [4] excludes clear activation of conscious access through the interaction between modules, because the architecture implicates a homogeneous mechanism that processes the content of various kinds from different functionally levels and the inputs. The same conclusions may come also from consideration of another architecture which postulates that mindreading and metacognition are parts of the one metacognitive mechanism [30–32]. In this account, it is believed that the attribution of mental states to other people depends on our direct access to these mental states (introspection) via subsequent processes of simulation and inference. Thus, human capacity for mindreading is based on the introspective data, which are initially accessed to imagine other's state, and then is used to make the attribution of this state to interpret or rationalize other's behavior (simulation).
