The text presents the activities of Dr. Kathleen Vaughan Wilkes, a philosopher from the University of Oxford in the Inter-University Centre Dubrovnik (IUC) from the beginning of the 1980s to the end of the millennium. Dr. Wilkes was co-directing the longest standing IUC course Philosophy of Science, but she also initiated other IUC academic programmes. As a member of the IUC governing bodies, she was highly engaged in securing scholarships for participants from Central and East Europe in IUC programmes, mostly through Open Society Foundation. Dr. Wilkes played a crucial role in spreading information from the city of Dubrovnik during the attacks of the Yugoslav People's Army in 1991 and during Croatian’s struggle for independence, for which she was awarded honorary citizenship and posthumously one of the squares was named after her.

This is a personal memoir about the life, work and courage of Professor Kathleen Wilkes, a Fellow in Philosophy for 30 years at St Hilda’s College, Oxford University. The article traces—and sets out to explain—particularly her links to Dubrovnik and Croatia and the Inter–University Centre since 1981, and supported strongly through the 1980s and even during the 1990s, remaining on site during the cruel siege of the city when the IUC suffered a devastating fire. Three key aspects of her life are explored—her work as a significant philosopher of science; her outstanding courage and work in defending academic freedom widely over the East Central European region, and her warm personality and generous friendship. This is why she can be regarded as Dubrovnik’s Global Citizen, the IUC was only too ready and willing to host this conference in her honour.

Kathy Wilkes contributed to two books on Goal-directed Behaviour and Modelling the Mind based on interdisciplinary graduate classes at Oxford during the 1980s. In this article, I assess her contributions to those discussions. She championed the school of philosophers who prefer problem dissolution to problem-solution. She also addressed the problem of realism in psychology. But the contribution that has turned out to be most relevant to subsequent work was her idea that in modelling the mind, we might need to “use as structural elements synthetic cells, or things that behaved very like neurones.” I show how this idea has been developed in my own recent work with zoologist and neuroscientist, Raymond Noble, to become a possible physiological basis for the ability of organisms to choose between alternative actions, and so become active agents. I consider that this insight became her seminal contribution in this field.

The primary aim of this article is to find out what different linguists say about the role of intentions in the study and explanations of language change. I try to investigate if in the explanation of language change, “having an intention” does any explanatory work. If intentions play a role, how do they do it, at which point it is salutary to invoke them, and what do they contribute to the explanation of language change? My main claim is that speakers’ intentions have a role to play only on higher linguistic levels, i.e., in speakers’ communicative strategies. Since this is a celebration for Kathy Wilkes and her contribution to goal-directed behaviour, in the Concluding remarks I go back to her remarks on language and intentions and see how they fit my discussion in this paper.

Machine learning researchers distinguish between reinforcement learning and supervised learning and refer to reinforcement learning systems as “agents”. This paper vindicates the claim that systems trained by reinforcement learning are agents while those trained by supervised learning are not. Systems of both kinds satisfy Dretske’s criteria for agency, because they both learn to produce outputs selectively in response to inputs. However, reinforcement learning is sensitive to the instrumental value of outputs, giving rise to systems which exploit the effects of outputs on subsequent inputs to achieve good performance over episodes of interaction with their environments. Supervised learning systems, in contrast, merely learn to produce better outputs in response to individual inputs.

How do we understand other individuals’ actions? Answers to this question cluster around two extremes: either by ascribing to the observed individual mental states such as intentions, or without ascribing any mental states. Thus, action understanding is either full-blown mindreading, or not mindreading. An intermediate option is lacking, but would be desirable for interpreting some experimental findings. I provide this intermediate option: actions may be understood by ascribing to the observed individual proto-intentions. Unlike intentions, proto-intentions are subject to context-bound normative constraints, therefore being more widely available across development. Action understanding, when it consists in proto-intention ascription, can be a minimal form of mindreading.

In her already classical criticism of thought-experimenting, Kathy Wilkes points to superficialities in the most famous moral-political thought-experiments, taking the Ring of Gyges as her central example. Her critics defend the Ring by discussing possible variations in the scenario(s) imagined. I propose here that the debate points to a significant dual structure of thought experiments. Their initial presentation(s) mobilize the immediate, cognitively not very impressive imaginative and reflective efforts both of the proponent and the listener of the proposal. The further debate, like the one exemplified by Wilkes’s criticisms and some of the answers, appeals to a deeper, more rational variety of imagination and reasoning. I suggest that this duality is typical for moral and political thought experimenting in general, conjecture that it might be extended to the whole area of thought experimenting.

We try not just to reconcile but to “integrate” Cognitivism and Behaviorism by a theory of different forms of purposiveness in behavior and mind. This also implies a criticism of the Dual System theory and a claim on the strong interaction and integration of Sist1 (automatic) and Sist2 (deliberative), based on reasons, preferences, and decisions. We present a theory of different kinds of teleology. Mere “functions” of the behavior: finalism not represented in the mind of the agent, not “regulating” the behavior. Two kinds of teleological mental representations: true “Goals” in control-theory, cybernetic view, with “goal-driven” behavior (intentional action); vs. Expectations in Anticipatory Classifiers: a reactive but anticipatory device, explaining the “instrumental” (finalistic) nature of Skinner’s reinforcement learning. We present different kinds of Goals and goal processing and on this ground the theory of what “intentions” are. On such basis, we can discuss Kathy Wilkes’s hint about the necessarily linguistic formulation of “intentions”; with the hypothesis that her intuition is not correct for any kind on “intention” which may be represented in sensory-motor format, but correct for “volition” and our will-strength for socially influencing ourselves.