(See all posts in this series here.)
Intention is a type of memory. I argue that research on working memory reveals the dynamics of intention as embodying the agent’s control in action. This is a new perspective argued for in Chapters 3 and 4 of MoM.
Elizabeth Anscombe noted that intentional action is subject to a why-question, “Why are you doing that?” to which one can give a reason for action. The answer tracks the agent’s knowledge of their action. Yet what if an evil neuroscientist, kitted with a precise neural disruptor, manipulates our brains so that at the moment we are asked why we are doing something, we no longer have an intention to access? Put simply, losing track of what we intend, we forget what we are doing. Given this amnesia, our response to the question would be: Um, what am I doing? (See the Continuity of Practical Memory, MoM Chp. 4.2). The thought experiment suggests that intention has a memory function. Intention is a practical memory or memory for work.
Talk of “memory for work” intentionally evokes the psychological construct of working memory. I understand working memory to be a subpersonal architecture, so it is not attributed to a subject. Rather, it explains the subject’s remembering for action (work).
Working memory (WM) is underdiscussed in philosophy. In Plans and the Structure of Behavior, Miller, Pribram and Galanter introduced the concept of working memory as a posit performing a specific function:
When we have decided to execute some particular Plan, it is probably put into some special state of place where it can be remembered while it is being executed…. we should like to speak of the memory we use for the execution of our Plans as a kind of quick access, “working memory.” …When a Plan has been transferred into the working memory we recognize the special status of its incompleted parts by calling them “intentions.”
Reading this, the scales fell from my eyes. Working memory was no longer “boring”. I saw in the vast experimental work on working memory, dissections of the dynamics of an agent’s intention. Intention isn’t some static state. It is constantly updating in time with action.
A traditional WM architecture such as Baddeley and Hitch’s separates WM into a central executive and a set of dynamic stores. Briefly, think of the central executive as set when the agent forms a plan of action, and the stores as retaining targets relevant to action as they appear and must be retained over time. With each new intention, the central executive is reset, and with each relevant change in the agent’s world, the stores are updated with action relevant targets. This dynamic is the substratum of the dynamics of intention.
Let me give the example, what I call practical fine-tuning (Chp. 4.3). We begin with a general intention to act on a type of object, say on Os. No Os have appeared yet, so the plan can only be general, representing the O-type. On perceiving a specific O, we fine-tune our intention to act on this O or such Os. The shift from a general intention to act on Os to a fine-tuned intention to act on this/such Os as perceived involves a practical inference: from general intention, the content of perception, to a fine-tuned intention.
- I intend to act on Os
- I perceive this as a relevant O
- I intend to act on this O.
Given (3), one acts or is ready to act. Beneath the change in intention is a modification of working memory as new targets are registered and encoded in mind. For example, in salient working memory experiments, subjects are instructed that there will be a type of task target, say color that leads to a general intention. A concrete color will then be presented (this color, say red). The subject then updates their intention to target that specific color with that sample kept in the appropriate store (see my discussion of retro-cueing experiments, an important working memory paradigm; MoM Chp. 4.4). The development of intention over time as in the above schema is underwritten by changes in working memory.
There are a number of interesting philosophical consequences of linking intention to working memory. Given recent work arguing that working memory stores are often realized in sensory areas, the retention of a fine-tuned intention amounts to setting vigilance for task targets, a readiness to attend. Intentions are also revealed to be dynamic states whose setting of attention proceeds through such vigilance (Chp. 3.5).
Let me briefly remark on our distinct access to action (Chp. 4). Earlier, I divided control from automaticity to resolve an incoherence in assumptions about intentional action. Control is tied to what intentions represent. Since the contents of intention is dynamic, control is in that sense dynamic. This dynamic allows us to track our actions in a non-observational way.
I speak of access rather than knowledge to minimize controversy. We have direct, non-observational access to what we are doing because intention can inform directly inform report. When we say what we are doing, we access our plan, what we intend. The intention that informs report constitutes a type of attention to what one is doing. Where the actions are mental, this access is introspection (MoM, Chp. 7), and intention grounds a type of introspective attention (Chp. 4.6). The critical point is that because intention is constantly updated, such as in practical fine-tuning, then our distinctive access to action is grounded in this dynamic updating. Where this updating is a form of practical reasoning, then practical reasoning provides the basis of access that keeps time with action as the world changes. There are Anscombean resonances here, but I’m not trying to explicate Anscombe’s views. Rather, the brief description is driven by understanding working memory as explaining the dynamics of intention, and intention as the source of our awareness of what we are doing. This is secured when we understand the mnemonic character of intention.
This completes my summary of salient parts of the core theory. In the last two posts, I apply it to actions of general philosophical interest. The next post concerns implicit bias in attention.
