It is both exhilarating and a bit bracing to have one’s work scrutinized by such a talented collection of critics. I am grateful to Greg Currie, Tyler Doggett, Anna Ichino and Jennifer Nagel for their immensely thoughtful and insightful challenges.
Responding adequately to all of the issues that they raise would take many more pages than I have been allotted. So I will limit myself to those that seem to be of widest concern. Most of these challenges centre on the account of alief that I offer in ‘Alief and Belief’ and ‘Alief in Action (and Reaction.)’ I will begin with some general remarks about the notion, and then turn to some of the specific worries raised in the individual papers.
1. Between reason and reflex
The idea of alief was initially introduced to make sense of a familiar phenomenon. We often find ourselves in the following situation. Our beliefs and desires mandate pursuing behaviour B and abstaining from behaviour A, but we nonetheless find ourselves acting – or feeling a propensity to act – in A-like ways.1 These tendencies towards A-like behaviours are highly recalcitrant, persisting even in the face of our conscious reaffirmation that B-without-A is the course of action to which we are reflectively committed. It seems misleading to describe these A-like behaviours as fully intentional: their pursuit runs contrary to what our reflective commitments mandate. But it also seems wrong to describe them as fully reflexive or automatic: they are the kinds of behaviours that it makes sense – at least in principle – to attribute to a belief–desire pair, even though in this case (I argue) it would be mistaken to do so.
The alief papers make the simple claim that it is conceptually useful to have a single category for the family of mental states that accompany these propensities.2 In an important sense, all the rest is commentary. (Of course, commentary is where most of the action is – and we’ll get there soon enough.) But first I want to defend the main point, thereby responding to some of the big-picture issues that are raised by all three of the response papers. The main point is simply that in developing a relatively coarse-grained folk psychology – one that makes use of the notions of belief and desire to explain intentional behaviours, and of the notion of reflex to explain behaviours that consistently occur in response to stimuli in the absence of representational mediation – it is crucial to have an intermediate category. My contention is that it should be one with the properties that I have attributed to the notion that I call alief.
One can think of the spectrum from reflexive behaviour to reflective behaviour as varying along two parallel dimensions. The first is that whereas the connections between stimuli and responses at the reflexive level are completely fixed, the connections between beliefs and desires – and thus between beliefs and their associated prescribed responses – are fully combinatoric.3 Alief – which lies in the middle of this spectrum – is molar: it has, I have suggested, a characteristic R–A–B structure, whereby a particular representation (R) comes typically to be associated with a characteristic valence or evaluation or affective state (A) and with the activation of a behavioural repertoire (B).4 This connection may be innate – that is, due to one’s ancestors’ repeated experiences – or habitually acquired – that is, due to one’s own. The second – related – contrast is that whereas fully non-intentional action due to reflex proceeds without representational mediation,5 and fully intentional action due to belief/desire involves representational mediation that is rich enough to generate the combinatoric possibilities alluded to above, quasi-intentional action due to alief lies between them: the representational content that contributes to an alief is processed in a relatively shallow way; in particular, it is not fully integrated with other representational content that may be simultaneously triggered by features of the subject’s internal and external environment.
None of my commentators denies that there are actions prompted neither by reason nor reflex. Currie and Ichino readily admit the existence of what they call ‘Gendler cases … cases of behaviours not characterizable as entirely automatic, but also not easily thought of as actions motivated by beliefs and desires; they are often cases where habit or a quick and dirty evolved response subverts the promptings of reason’. Doggett has no qualms about presenting the wallet case as involving a kind of ‘absent-minded reaching’ that is neither fully reflective nor purely reflexive. And Nagel is happy to acknowledge that there is a ‘murky zone where the mind lumbers forward on autopilot, and the line between genuine action and mere bodily motion seem strangely blurred’.
Where Currie and Ichino, Doggett – and to some extent Nagel – disagree with me is about the proper explanation – or characterization – of these phenomena. In the next section, I will try to motivate my account in ways that directly address their concerns.
2. (Not) kicking the can
Rather than put off for later what can easily be done now, let’s look at a specific set of cases. Inspired by Nagel’s opening pair of sentences, I’ll begin with a set of examples familiar from action theory. In each of them, the story starts with me seated normally in a chair with my legs bent and my feet on the floor, and finishes with my left leg extended so that my foot hits against the trashcan, knocking it over. The first three versions run as follows:
Belief–desire: I believe that extending my leg and foot will cause me to kick over the trashcan, and I desire to kick over the trashcan. I extend my leg and foot, and kick over the trashcan.
Reflex: You strike my knee with a tendon hammer just below the patella. This causes my leg and foot to assume the extended position, thereby kicking over the trashcan.
External compulsion: My leg and foot rest on a specially constructed hydraulic lift. You push a button, which causes the lift to raise my leg and foot to the specified flexed position, thereby kicking over the trashcan.
Case (1) describes a paradigmatically intentional, voluntary action: it is goal-directed and rationalized by a belief–desire pair that (let’s posit) I recognize and endorse even after deep reflection. Case (3) is not even an action. What about Case (2)? There, my kicking over the trashcan results from what Aristotle would call ‘an external principle, the sort of principle to which the agent, or [rather] the victim, contributes nothing’ (Aristotle, NE III.1, 1110a3–5): the action is neither voluntary nor intentional.
There has been – to put it mildly – a rather sizable discussion of these issues in the last 2000 years.6 Among the many important strands is one that has focused on the ways in which compulsion or ignorance may interfere with voluntary or intentional action in cases like (1).7 Suppose I develop my desire to kick over the trashcan only because you threaten to harm me if I do not. Here, it seems reasonable to say that my kicking of it – though perhaps intentional – was not voluntary. Suppose (to change the example slightly) that what I thought was a trashcan is actually a bucket of water in disguise. Here, it seems reasonable to say (de dicto) that my kicking of the bucket was not either intentional or voluntary.8 But though these are cases where my action – at least under a certain description – is not voluntary or intentional, they are still cases where my action is governed by a belief–desire pair: it’s just that the desire (in the compulsion case) or the belief (in the ignorance case) is in some way deficient.
The cases that interest me in the context of the alief papers are deviant along another axis. Suppose that I am playing an odd sort of game in my office that has the following rules: whenever I see a trashcan, I should kick it as fast as I can, unless it is between 12:15 and 12:20. Suppose that it is now 12:17, and that I am occurently aware both of that fact, and of the time-indexed no-kicking rule. But suppose that because I have become so trigger-happy in my kicking, I extend my leg regardless.
Habit: I believe that extending my leg will cause me to kick over the trashcan, but I don’t desire to kick over the trashcan. Out of habit, I extend my leg, thereby kicking over the trashcan.
Or suppose that I am playing the game in my office, and I know that you are experimenting with the new holographic projector. And suppose that I don’t want to waste my energy on non-productive kicking, since that will exhaust me needlessly. Suppose that you now project a holographic image of a trashcan before me, and that I am occurently aware both of the fact that I am gazing towards a mere image, and of my desire to kick all and only actual trashcans. But suppose that because I have become so trigger-happy in my kicking, I extend my leg regardless.
Misleading perceptual stimulus: I don’t believe that extending my leg will cause me to kick over the trashcan, and I desire to kick only if doing so will enable me to kick over the trashcan. Aware that my efforts will be futile, I extend my leg, thereby kicking the air.
How should we describe my kicking of the (apparent) trashcan in Cases (1.5) and (1.6)? Unlike Cases (2) and (3), where my mental state makes no contribution to the extension of my leg, the (1.x) cases are cases where the representational content plays a role in triggering my action. It’s just that – as I noted in Section 1 – the representational content is processed in a relatively shallow way; in particular, it is not fully integrated with other representational content that has been simultaneously triggered by features of my internal and external environment. And because this content happens to be associated with a particular motor routine – the extension of my leg in the direction of the apparent trashcan – that routine is readied and executed, even in the face of my explicit belief/desire-driven decision to pursue a contrary course of behaviour.
Obviously, Cases (1.5) and (1.6) are a bit contrived. But they are – at their core – like the belief-discordant alief cases that I present in the original papers. Broadly understood, each of the cases in the papers falls into one of these two categories. Either – as in Case (1.6) – I am presented with a misleading perceptual (or cognitive) stimulus that I recognize as such, but toward which I nonetheless (feel the propensity to) respond in my typical manner. Or – as in Case (1.5) – I am presented with an accurate perceptual (or cognitive) stimulus towards which I have formed the (local or standing) intention to respond in a particular way, but towards which I nonetheless (feel the propensity to) respond in some other manner. Though Cases (1.5) and (1.6) involve propensities towards motor routines, the propensity in question may instead be towards the rendering occurrent of a particular mental state.
At the risk of disappointing my readers who have been under the impression that I was on to something radical: that’s it. That’s all I mean by alief. And – indeed – that’s all I ever said it was. ‘An alief is, to a reasonable approximation, an innate or habitual propensity to respond to an apparent stimulus in a particular way’ (Gendler 2010: 282; 288). ‘A paradigmatic alief is a mental state with associatively linked content that is representational, affective and behavioural, and that is activated – consciously or non-consciously – by features of the subject’s internal or ambient environment’ (Gendler 2010, 263ff). All the rest is commentary.
But – as I said at the outset – commentary is where most of the action is. So let me turn to that.
3. The cases
Let’s begin with the wallet case, which is a focus of much of Doggett’s discussion. I think – and I think Doggett also thinks – that this is akin to Case (1.5). Here, I am presented with an accurate perceptual stimulus (the money) towards which I have formed an intention to respond in a particular way (by putting it somewhere other than in my (absent) wallet), but towards which I nonetheless feel the propensity to respond in my habitual manner.
Turn now to the cluster of cases that Currie and Ichino enumerate in their footnote 4 (2012: 791) – though they do so with the aim of casting aspersions on the alief account. They list there the cases of ‘hurrying when looking at your watch, deliberately set five minutes fast (261; 286); reluctance to drink from a sterilized bedpan, to eat fudge shaped like dog faeces, or to put in your mouth a clean piece of vomit-shaped rubber (257; 262; 266); a frog’s snapping at an inedible small black moving object (283; 286)’
I think – and I think Currie and Ichino might agree – that all of these are akin to Case (1.6). They involve a misleading perceptual stimulus to which the subject (feels a propensity to) respond in a particular way as the result of the activation of a locally inflexible R–A–B cluster. Basically, if R typically occurs when I am in situation S9 – or is somehow relevantly associated with my being in situation S – then the following may occur even when I know (or believe) that I am in a non-S situation:
Misleading perceptual stimulus: I don’t believe that I am in situation S (indeed: I believe that I am not in situation S). And I have no independent desire to behave in the ways that I typically do in situation S. Nonetheless, I (feel a propensity to) act in ways that I typically do in situation S. This is because I am experiencing an S-typical R, which activates a locally inflexible R–A–B structure.
To be fair, this is a bit too quick, and brings out a way in which I have been simplifying a problem that is a good deal more complex than I have been letting on. As a number of Nagel’s remarks nicely bring out, I have been insufficiently candid about the complex interplay between internal and external environment in determining which R–A–B structure is activated by (what we naturally describe as) the salient stimulus. Seeing 10:05 on my fast-set watch will trigger one sort of A–B sequence on a day when I teach at 11:00, and quite another on a day when I teach at 10:00.10 Smelling a rich and savoury dinner may trigger one sort of A–B sequence when I am hungry, and quite another when I am sated to the point of nausea.11 How is this possible? It’s possible because – speaking loosely – we typically have multiple aliefs which may be triggered by a given stimulus. Which one is activated on a particular occasion will depend on context. (If we are speaking strictly, of course, we can build this into the R.)
For all that, I don’t think this is where our deep disagreements lie. I think they lie with the next concern, which I call ‘the hodgepodge worry’.
4. The hodgepodge worry
Both Doggett and, Currie and Ichino raise a worry about whether alief is a fundamental mental state – as opposed ‘an amalgam of several more primitive mental states: those of entertaining content R, experiencing affect A, and activating behavioural repertoire B’ (Gendler 2010: 264, note 16, emphasis added12). Doggett does a nice job of articulating the concern. He writes:
And, he continues: ‘Is there a unity to alief where this goes beyond the fact that typically there is the co-activation of alief’s components? If so, is the unity doing explanatory work for Gendler? What?’
Even if Gendler is right about the identity of the representational state and about co-activation, her view doesn’t yet follow. Her view is not simply that there is a representation and it produces some affect and some behaviour preparation. Rather, it is that there is a state comprising these three bits, and, furthermore, that state is not simply a hodgepodge of the three. To use some imagery from Leibniz, alief is sheep rather than flock. To use some imagery from David Lewis, it’s trout rather than trout-turkey
Currie and Ichino raise a similar concern in slightly different form. They point out – correctly – that when ‘Humeans say that it is always a belief–desire pair that motivates; they are not thereby committed to saying that action is produced by composite entities with beliefs and desires as their parts’. They write: ‘When people believe they are being pursued by a bull, they almost invariably fear the bull, (occurently) desire not to be caught on its horns, and try to escape. Each of these states is rarely activated except in combination with the others. But we don’t give a better explanation of bull-avoiding behaviour by saying that the belief, the desire and the emotion are components of a single entity.’ That is, like Doggett, they challenge me to explain why the R–A–B components of an alief form a single composite entity, while the belief–desire-action components of an intentional action do not.
The explanation, I think, is not hard to come by. On my view, the reason it is misleading to speak of a belief–desire pair (or belief–desire–action trio) as a composite entity is because the relation between my beliefs and my desires is – at least in principle – fully combinatoric. This is true even if – as in Currie and Ichino’s example – the belief that a bull is charging is typically associated with a desire for quick retreat.13
Here is a simple test to determine whether my running from the bull is an intentional action, generated by a belief (that the bull is charging) and a desire (to retreat if a bull is charging) – as opposed to being a quasi-intentional action driven by a relatively unified state such as alief. Is there additional information about the bull’s charging that could change my desire, and thereby change my action-propensity? (Suppose, for instance, that I were to learn that being gored by the bull would actually be a painless path to eternal salvation – and that learning this would cause me to have the desire to be gored by the bull, thereby rationalizing an action of running as quickly as I can towards its horns.) If so, then there’s no reason to speak of the belief–desire pair as a composite entity, and every reason to speak of my retreat as fully intentional – since the association between the belief and the desire is (psychologically) contingent.
I suspect, however, that elements of my bull-retreat behaviour are traceable to alief rather than belief – though this may be invisible in a case such as this one, when (as a matter of fact) the alief-mandated behaviour and the belief-mandated behaviour coincide. The test is in the counterfactual: Even if reason were to tell me to stand firm, would I still cringe – and perhaps even retreat? If so, the bull-charging case would turn out to be what Currie and Ichino call a Gendler case: one where ‘habit or a quick and dirty evolved response’ would ‘subvert the promptings of reason’. The same response – mutatis mutandis – can be offered back to Doggett, whose target here is the wallet case.
5. But why alief?
Fair enough, you might say: but why do I think that this shows that it is useful to have a notion of alief? Why is it not enough for me simply to say that habit plays a much larger role in our lives than we are (pre-reflectively) inclined to admit? Well, at some level, what it is to say that it is useful to have a notion of alief is to say that habit plays a much larger role in our lives than we are (pre-reflectively) inclined to admit. But note that this is ‘habit’ in a very special sense of the term. It has to have the characteristically R–A–B structure that I have identified in my characterization of alief. The difference between habits (narrowly construed) and innate propensities has to be immaterial in this context. And appeal to habit has to be a helpful way to explain a wide range of behaviours that are typically taken as indicative of a subject’s beliefs. That is: as long as you mean by habit what I mean by alief, I have no quarrel with this paraphrase.
Except that I do. And that’s because if one takes the notion of alief seriously, there’s a reason to think that labels matter.
When hospitals want to reduce the rate of errors in the operating room, one of the first changes they make is to ask all of the participants – doctors, nurses, orderlies – to call one another by their first names. Why? Because when a nurse calls the surgeon ‘Dr Such-and-Such’, this activates in both of them a particular sort of R–A–B, through which their subsequent interactions are filtered.14 Patterns of attention, encoding of information, interpretation of ambiguities – all of these are subject to the vast and fascinating set of subtle top-down and bottom-up patterns of interaction that cognitive and social psychology have been studying for the last century.15 The categories we have for making sense of our experiences elicit certain ‘habits of thought’ – association-driven patterns of attention, encoding and interpretation.
So here’s a reason not to simply describe Gendler cases in terms of habits: because our habitual associations with the term ‘habit’ are too narrow. Its paradigm cases lie too close to one end of its range.
Part of the point of lumping together such an apparently motley collection of cases under the notion of alief is to bring out just how large a domain I mean to be accounting for. (For a full reckoning, see Section 7 below.) Part of the point is to elicit – in an alief-like way – thoughts of the skywalk-walker and the bedpan-drinker and the watch-setter and the wallet-depositer and the movie-goer and the kitchen-re-arranger and the sports-shouter and the race-associator any time you start thinking about any one on them. Part of the point of the alief papers is to make you alieve in alief!
6. Loose ends and quick concessions
In the limited space allotted to this response, I have left many important worries unaddressed. As any reader of my insightful critics recognizes, to answer these adequately would require many pages. But in some cases, I am prepared simply to plead guilty as charged. (The remarks in this section are quick, and are likely to make sense only if you have had a chance to read through the three critical essays to which they are responses.)
6.1 Currie and Ichino
Currie and Ichino are – I suspect – correct in their claim that my hyperopacity argument rests on an ambiguity. Alief is – of course – hyperopaque in their sense H1.16 But they have convinced me that a good deal more work is required for me to show that this distinguishes alief from imagination – or, for that matter, from any other mental state.
I am also intrigued by their tantalizing introduction to alief’s ‘unpredictable and novel … wildly promiscuous … exciting and disreputable’ cousins. Their basic idea here – that ‘both creative and habitual processes are triggered by an initial representation – not always specifiable at the personal level and sometimes occurring beyond conscious will – that leads, through a chain of barely conscious associations, to various other states, the last of which can be either another representative state or an action’ – strikes me as an enormously promising avenue to pursue. I very much hope that they – or I, or you – will find the time to follow through with this.
Doggett is – I think – correct in his contention that not every vivid action-generating imagining is an act of self-deception. This means, as he rightly points out, that some crucial element is missing from the account that I offer in ‘Self-Deception as Pretense’. Doggett suggests that this ingredient is that self-deceived subjects often have false beliefs about what they believe. This strikes me as a helpful amendment. Self-deception is one of a family of cases where one’s thoughts and actions (in some particular domain) are guided in a reality-indifferent fashion. What distinguishes self-deception from some of the other cases in this family is that it involves a partially but not fully extended lack of self-awareness about what one is doing.17 But though I am ready to concede that the Pretence account requires emendation, I’m not ready to go all the way to what Doggett calls ‘Self-Deception as Alief’. I find the idea an intriguing one, but I need to give it further thought.
Nagel is absolutely correct that there are cases – the Iowa Gambling Task is one, but there are others – where our implicit mechanisms for statistical encoding of patterns are ‘better able to keep pace with variation in the world’ than are our considered judgements.18 But I’m not ready to classify these cases as beliefs (see Section 6.4 below.) They are akin to the implicit attitudes – such those measured by the IAT – that I take to be some of alief’s paradigm instances. So I need to do more thinking here. So too with the interesting Fazio studies she describes where the implicit attitude appears to be more malleable than the explicit. As for spider phobia: my understanding is that the empirical evidence is somewhat mixed, but if things are indeed as Nagel reports, then she is right that an alief approach to these cases will be somewhat complicated.
On the other hand, I take as a friendly suggestion her recommendation that I think more carefully about the relation between my alief view and the Gawronski group’s APE model.19
6.4 All three
Finally, all my critics are absolutely correct in pointing out crucial inadequacies in the account of belief to which I tacitly appeal.20 Here, I can say only that if there were a ready-to-wear characterization of belief that fully fit my needs, I would happily purchase it. My failure to find one in the marketplace of ideas only reinforces my sense that the set of skills required for this tailoring project far exceed my own.
7. The bigger picture
For all that I have said above about the ways in which the alief account is anything but radical, I don’t want to be coy about how much it is that I am claiming. Although I motivate the notion of alief by pointing out its utility in accounting for a large class of belief-discordant behaviours, I actually think alief is what governs most of our actions, most of the time. Nagel notes this in her closing paragraphs. She writes: ‘[Gendler’s] theory is motivated by a close study of cases in which there is some tension between our behaviour and our reflective attitudes, but Gendler suggests that its application may in the end be much wider: in her view even the behaviour that aligns with what we reflectively endorse could typically be driven by the more primitive state of alief (266, 281).’21
Nagel notes that ‘this idea runs against our usual self-conception’ – and this I do not deny. But my recent forays into the field of empirical psychology have rendered me particularly susceptible (or, as Nagel puts it ‘uncommonly alive’) ‘to the possibility we are not what we seem to ourselves to be’. I am, in other words, particularly susceptible to the Socratic idea that we are – in a very profound sense – opaque to ourselves, and that one of the tasks of philosophy is to help us become aware of the ways in which – in the words of Henri Amiel – ‘Our greatest illusion is to believe that we are what we think ourselves to be’.22