Hasan Çagatay

Understanding is the foundational component of intelligence that makes adaptation to novel states possible. As artificial intelligence models advance towards artificial general and super intelligence, an objective, in contrast to human-centered, theory of understanding is required. One merit of an objective conceptual analysis of understanding is that it would unify human and artificial understanding. This study defines understanding as well-integrated use of reasoning across a network of conceptual, logical, and causal relationships, which are situated on an efficient conceptual web. The evolutionary function of understanding is the capacity to generalize knowledge into novel situations. This renders machine learning concept “generalization” a measure of understanding. The paper discusses two primary methods for assessing artificial understanding: behavioral methods measuring generalizability and mechanistic analyses revealing causal mechanisms of the neural network. We discuss the limitations of these methods. The most important problem, as this paper defends, is the difficulty of measuring objective understanding in contrast to human-understanding.

This paper aims to contribute to the ongoing discussion on adherence to social norms. It considers insights from multiple research traditions in an effort to explain how individual learning and action are connected to social norms. One strand of philosophical tradition holds that non-representational learning and skillful coping carried out unconsciously are underestimated by both scientific and philosophical traditions. The present research combines this tradition with the literature on the evolution of social norms and suggests that experienced individuals in a society adhere to social norms better than novice agents do. We explain this phenomenon by unconscious and non-representational cognitive processes. This framework is then used to investigate population-level outcomes of individual learning.

A famous puzzle called “Grandmother Paradox” is used to argue against the feasibility of traveling backward in time because of the logical and nomological problems such travel involves, and not only because we don’t have the technology to make it reality. The same kind of problems would be encountered in leaping forward in time and then returning to the time of departure. We argue that a similar family of problems also arise in our having foreknowledge of the future without making any time travel. We point to the mysterious consequences of having access to a being, say a machine or a psychic, that can have infallible knowledge of the future and conveys this foreknowledge to human beings truthfully, without any lies or distortions. The cause of these mysterious consequences is the fact that such machines or psychics will raise logical and nomological complications reminiscent of the ones we encounter in time travel scenarios, and that is a strong reason why infallible foretellers cannot exist. We conclude that we can have foreknowledge of the future, in principle, only within certain narrow limits, if at all.

In 1980s, neuroscientists joined philosophers and psychologists in the investigation of volitional actions and freedom of will. In a series of experiments pioneered by Benjamin Libet, it was observed that some neural activities correlated with volitional action regularly precedes the conscious will to perform it, which suggests that what appears to be a free action may actually be predetermined by some neural activities, even before the conscious intention to act arises. Shortly after publication of that study, Libet’s findings and interpretations were started to be criticized on philosophical and methodological grounds. In this study, the legitimacy of the criticisms directed to Libet’s and his successors’ experiments is discussed by taking recent neuroscience studies on volition into account and it is argued that these criticisms are not sufficient to eliminate the doubt that these experiments casted on the freedom of the will.

By the Chinese room thought experiment, John Searle advocates the thesis that it is impossible for computers to think in the same way that human beings do. This article intends firstly to show that the Chinese room does not justify or even test this thesis and secondly to describe exactly how the person in the Chinese room can learn Chinese. Regarding this learning process, Searle ignores the relevance of an individual’s pattern recognition capacity for understanding. To counter Searle’s claim, this paper, via examining a series of thought experiments inspired by the Chinese room, aims to underline the importance of pattern recognition for understanding to emerge.