1. Do I Exist?

It is an extraordinary thing that Descartes' famous Cogito argument is still being puzzled over. For over three hundred years philosophers have argued about how long the Cogito argument is, about how many parts it has, about what it aims to do, and about whether it works. This paper is another fragment in that untiring tradition of puzzlement.

Let us assume that the Cogito seeks to answer the question ‘do I exist?’ If I were, for any reason, looking for re-assurance about my own existence and were thereby led to ask the question ‘do I have a grounds for thinking I exist?’ would the Cogito furnish me with a positive answer? I am going to argue that the Cogito can be construed in such a way that it does provide for a positive answer to that question. In the Second Meditation, Descartes engages in thought, judges ‘I am thinking’ (cogito), and from that, rightly in my view, infers ‘I exist’ (sum). We can do the same. On my understanding of it, the Cogito is an argument type that enables each of us to establish our own existence simply on the basis of our own conscious acts of thinking. A subject who engages in conscious thought, judges (on that basis) that they are thinking, may rightly infer on that basis that they exist.

John Campbell in his ‘Lichtenberg and the Cogito’ argues in favour of Cogito-scepticism. Campbell claims that the Cogito is either too weak to provide us with an existential conclusion, or it is question begging.Footnote ² Either way, the grounds of the Cogito do not, on his view, provide a subject with a reason to judge she exists.

My aim in this is to argue, against the Cogito-sceptic, that there is a way of construing conscious thinking on which the Cogito can be seen to provide a non-question begging argument for one's own existence.

2. Campbell's Objection to the Cogito

To start let us set out the target argument in the way that Campbell does, and state his objection. The target argument runs as follows:

So understood, the Cogito argument has three parts. The first part is not a premise or a judgment: it is an occurrence. In particular, it is an engagement in a particular act of conscious thinking. The second part of the argument is a judgment: the judgment ‘I am thinking’. This judgment is supposed to be grounded in, but not inferred from, the first part – the engagement in a particular act of conscious thinking. The third part of the argument is the conclusion we are aiming at: the judgment ‘I exist’. The judgment ‘I exist’ follows by inference from the judgment ‘I am thinking’.

The target argument according to Campbell faces a dilemma: either we get to the conclusion by begging the question, or we do not get to conclusion.

Let us consider the ‘either’ fork first. The claim, in essence, is that (3) needs to be assumed to get from (1) to (2): you need to have knowledge of your own existence i.e. knowledge of the conclusion ‘I exist’ in order to be able to move from engagement in the particular act of conscious thinking to the judgement that ‘I am thinking’. The grounds for the claim that engagement in conscious acts of thinking are insufficient to ground judgements about thinkers are to a large extent the traditional Lichtenbergian grounds: ‘“Thinking is going on” is what one should say, just as one says “Lightning is occurring”. Saying “Cogito” is too much, as soon as one translates it as “I am thinking”’Footnote ⁴

How do I know merely from a particular act of conscious thinking that I am thinking? Maybe all I can know is that there is thinking going on? In order for me to know that the thinking occurrences are being had me, do I not already need to have some reason, either independent of the fact that I am thinking, or invoked by it, for believing that I exist? And if I need already to have these reasons, then I can get my transition between (1) and (2), but I have begged the question because I am using independent grounds for (3) to make the transition. Thus, the argument needs to assume what it seeks to establish to get from the first step – the act of conscious thinking – to the second. If the use of ‘I’ in (2) implies that a self exists, then you need to know you exist before you are entitled to use it to report your conscious thinking and so, as Campbell explains it:

Now let us turn to the ‘or’ fork. The claim on this fork of the dilemma is that while there may be a way to get from (1) to (2), without assuming (3), it is a way that then does not allow us to get to (3).

Suppose we claim that we can ground the judgment ‘I am thinking’ in an engagement in a conscious act of thinking without already assuming that ‘I exist’ because uses of ‘I’ need not carry referential import. Rather in judging that ‘I am thinking’ I am operating with a use of ‘I’ that has merely perspectival import. To illustrate the possibility of a notion having perspectival, but not referential import, Campbell looks to the case of temporal notions. I may not realize that there are time zones when I identify the current time as ‘5 o'clock’, but I am doing so relative to the time zone I occupy. It does not follow, he argues, that in holding that it is ‘5 o'clock’ I am referring to that time zone.Footnote ⁶ The suggestion is that we may use ‘I’ in a way that is relative to the person I am but does not refer to the person I am. The thought seems to be that I may use ‘I’ in a way that is relative to the subject I am, even when I do not realize I am a one subject rather than another. And if that is so then we can get from (1) to (2) without assuming (3). However, we then face a problem with the move from (2) to (3). If we construe the ‘I am thinking’ non-referentially we are not then entitled to move from ‘I am thinking’ to ‘I exist’. Using the arrow diagram used by Campbell we can represent the situation as follows:

1. (1) a particular conscious thinking

   No assumption that I exist

2. (2) I am thinking

   

3. (3) I exist

3. Can We Avoid Begging the Question?

I think it is clear that if we withdraw to a use of ‘I’ which is non-referential, then there is no getting from (2) the judgment ‘I am thinking to (3) the judgment ‘I exist’, unless we have a similarly ‘non-existential’ notion of existence – and what could that be? So, our only hope if we want maintain the claims of the Cogito to provide us with a way of gaining knowledge of our existence is to challenge the question-begging charge. In particular we need a way to challenge the claim that an engagement in a conscious act of thinking is not sufficient to warrant a subject in moving to the judgment ‘I am thinking’, unless she independently assumes her own existence.

Campbell compares the Cogito to Moore's famous ‘Proof the External World’ and draws out interesting parallels between the two.Footnote ⁷ Campbell construes Moore's argument as follows:

Both arguments seem to have three components, the first of which is a psychological conscious occurrence, the second of which is an indexical or demonstrative judgment, and third of which is an existential judgment. In Moore's argument we have a conscious visual perception that grounds without inference the indexical judgment ‘this is one hand, and this is another hand’, from which we can infer the existential judgment that ‘external objects exist’. In the Cogito argument we have a conscious act of thinking that grounds the indexical judgment ‘I am thinking’, from which we can infer the existential judgment that ‘I exist’.

It is commonly objected that Moore's argument fails because we need to assume knowledge of the existence of external objects in order to be warranted in moving from (1b) to (2b). Now Moore's argument was famously thought to involve a problem of a similar kind to the Lichtenbergian problem that Campbell puts to the Cogito argument, namely that it is acceptable to move from (1b) to (2b), only when we have assumed or independently establish (3b) – which was supposed to be our conclusion. So again we're faced with the dilemma that, either we get to the conclusion by begging the question, or we don't get to the conclusion.

The familiar complaint against Moore's argument is usually fleshed out via an argument from illusion. Suppose I have a visual perception of my hands, and I judge on that basis that ‘this is one hand, and this is the other hand’, it may be objected that you cannot justifiably move to the conclusion that ‘external objects exist’, unless you assume that your visual experience was caused by your hands. After all, it is objected, if you had been hallucinating, or be subject to an illusion, you could have had the same visual experience and it not be caused by your hands. If that is true then it looks as though you are going to have to assume existence of external objects, alongside with your visual perception, in order to move to your conclusion that external objects exist – and that is begging the question. If instead we try to row back from the assumption that external objects exist, and construe the judgment ‘this is one hand, and this is another hand’ in such a way that can be grounded on the visual perception alone, we will not, the argument goes, have sufficient resources to reach an existential conclusion. When I judge that ‘this is one hand, and this is another hand’ I am not thereby referring to any external object. My uses of ‘this’ are used with perspectival, but non-referential, import and used properly across veridical and illusory cases.

So, we see a parallel objection to both the Cogito and to Moore's argument. In relation to both we can object that a conscious experience is not itself able to deliver up knowledge of the existence of objects: selves or hands. To draw our conclusions we need already to know that there is some object or self beyond the experience; we need to know that it is not a mere experience of nothing, had by nothing. But to rely on such knowledge would be to beg the question.

Despite so elegantly bringing out the parallels between the two arguments Campbell's central concern is in fact to claim a contrast between them. Campbell rehearses an increasingly popular defence of Moore's argument, but claims that a parallel response is not available to the defender of the Cogito. It is that claim I want to examine.

What is the popular defence of Moore's argument against the charge of question begging? It is to point out that while it is true that on certain ways of construing the nature of visual perception one would need to add a further assumption (that, say, an external object is the causal source, or the represented object, of the perceptual experience) in order to justifiably draw the conclusion that external objects exist, such a construal is not compulsory. There are other ways to construe the nature of visual experiences on which such an added assumption would not be required. Suppose, we take a relationalist, or direct realist view of visual experiences. On such a view to have a visual perception of a hand is to have an experience in virtue of standing in a direct relation to a hand. If there were no hand, there would be no such visual experience. Therefore, if we are actually seeing one hand and then another hand, that visual experience can by itself function as grounds for the judgment ‘this is one hand, this is another hand’ without any independent premise being required. So if, as Campbell puts it, it is possible to argue that the external object is not in fact ‘beyond’ or ‘external to’ the visual perception, but rather the visual experience ‘encompass[es] the external object’, then the visual experience will be ‘enough on its own to generate knowledge of the existence of external objects’.Footnote ⁹

Having rehearsed the resources of a relationalist response to the question-begging challenge facing Moore's argument, Campbell makes this intriguing remark: ‘in contrast, in case of the Cogito there seems no possibility of a disjunctive or relational understanding of your relation with your thought on which your encounter with the thought encompasses not just the thought but the thinker’.Footnote ¹⁰

For the remainder of this paper I want explore whether we should accept that there is no such possibility. Could we not construe what it is to be engaged in a conscious act of thinking in such a way that it ‘encompass[es] me’ and is ‘enough on its own to generate knowledge’ that I exist?

4. Solvitur Ambulando?

I want to suggest that the prospects for the view that in being engaged in a conscious act of thinking a subject has a direct experience of herself of a kind sufficient in itself to ground the judgment ‘I am thinking’ might be made more evident if we think about our experience of acts and activities other than thinking.

Let us, for example, think about the conscious activities of walking or jumping. Walking and jumping are things that I do. They are also things I am aware of doing as I do them.

The first question to ask about our awareness of such activities is whether we have any reason to be more skeptical of having a direct awareness of them, than that we have direct awareness of our hands or coffee cups. When I am conscious of my walking or my jumping, my walking and my jumping seems to be as immediate and directly accessible to me as anything given in perception. And when I am conscious of your walking or jumping, your walking and jumping seems to be as immediate and directly accessible to me as anything else given in perception.

The second question to ask is how we should understand the relation that holds between our awareness of the actions and activities, and our awareness of the agents that carry them out. Let us suppose you have an awareness of my walking. You see me walk across the room, for example. We would think it very odd if you were to claim that while you saw my walk you could not, or did not, see me. Rather, when you see me walking what you see is me doing something: you see me in a certain mode, carrying out a certain set of bodily movements that are my walking.

The thought is that if you put the answers to the two questions just asked together, then there is scope to claim that we can be directly aware of agents in being directly aware of their actions. If we are aware of activities or actions by being aware of an agent doing something, and we are directly aware of those actions then the agent would seem to be a candidate for being ‘encompassed’ within the experience of the activity.

If it turns out that there is no insurmountable impediment to my understanding my awareness of my action (walking or jumping) as a direct relation to these activities, and that the relationship between actions and activities and the agents that carry them out is such that if you are aware of the action or activity you are aware of a mode of an agent, then we begin to have the resources to mount a response, of a kind the relationalist about visual perception mounts against the charge that Moore's argument begs the question, to the charge that the Cogito begs the question.

There are three claims made by the view being mooted. The third is supposed to follow from a proper understanding of the first two. The three claims are:

1. 1. A direct awareness of action thesis: when we consciously act we are directly aware of the activity/action. When we are conscious of others acting we are directly aware of the activity/action.

2. 2. An activities and actions as modes of an agent thesis: activities or actions are dependent on agents, in virtue of being modes of agents – they are ways an agent is being, or has been.

3. 3. A direct awareness of agent thesis: to be directly conscious of an activity/action is to be directly conscious of a way an agent is, and so directly conscious of an agent. (For example, to be directly conscious of a jumping is to be conscious of the jumper jumping, to be directly conscious of a walking is to be conscious of the walker walking.)

Let us suppose that claims 1–3 are true of our awareness of walking. If they are then we have reason to think we have available to us an argument that is capable of being used to establish our own existence that does not fall foul of either the insufficiency charge, or the charge of question begging. We can call the argument the Ambulo argument. Its structure is similar to that of the Cogito, and of Moore's argument, and comes in three parts: A psychological occurrence, an indexical judgment, and an existential claim:

The Ambulo – assuming theses 1–3 are true of a subject engaged in consciously walking – is successful in grounding in a non-question begging way the judgment ‘I am walking’.

It is able to do so in the same way that Moore's argument is able to ground the judgment that external objects exist in our direct awareness of them. We have construed the experience of acting is such a way that the acting, and so the actor, is not beyond or external to the conscious experience, but is ‘encompassed’ in it. In consciously walking I'm directly aware of the walking and thereby of the walker, and can on that basis infer that I exist. And of course if the Ambulo works as a proof of my existence then there is shed more where that came from. For example:

Let us then turn back to the intricacies of the Cogito. If they can be made to work with the right assumptions about the nature of our experience of activities and their relations to agents, and if thinking is rightly understood as an activity of a subject – along the lines that walking and jumping are – might we in fact have a non-question begging Cogito argument. Suppose, as well 1–3, we also claim:

When we then go back to the Cogito we are able to see a form of argument which would, if what seems to be true for walking is true of thinking, provide us with conception of conscious activity that is able to give us direct awareness of ourselves if we are engaged in such an activity – and in doing so can ground the judgment, I am thinking, and in term the judgement ‘I exist’:

5. Gassendi's Ambulo

The idea that we should look to the Ambulo argument alongside the Cogito argument to throw light on the latter is not a new one. It is already there in Gassendi's objections to Descartes. Gassendi points out that there is nothing very special in the form of argument provided by the Cogito and claims that Descartes ‘could have made the same inference from any one of [his] actions, since it is known by the natural light that whatever acts exists’.Footnote ¹¹

Descartes' replies to Gassendi as follows:

For our purposes there are three things to note about this exchange. First, Descartes' objection parallels the standard objection to Moore's argument. He points out that we cannot know for certain that we are walking because we have erroneously had the experience of walking even when we are not – for example, when we are dreaming. This fact is supposed to undermine the possibility of the Ambulo giving us knowledge that we exist. Descartes assumes that our experience of walking is not to be construed as essentially involving walking. But if we were to adopt relationalism about our experience of walking we could deny this. Second, Descartes does not deny that ‘it is known by the natural light that whatever acts exists’ and does not dispute the idea that if you know the act you know that the actor exists. The third thing to note about Descartes' response to the argument is that his concern is to emphasize that in the case of walking, in contrast to thinking, you don't know the act with certainty. You might think you are walking – as in a dream – but you might be wrong. So, what thinking is supposed to give us is certainty. The point here is that if walking is activity of mine, then nothing has been said to block a non-question begging argument for my existence. The conditional nature of that claim should be noted here. If walking is not an activity of mine that picks out a way I am, then being aware of walking will not be a way of being aware of myself. And Descartes might indeed be skeptic about walking as an action of the subject. It is a delicate question whether Descartes takes the subject of my bodily activities to be me – if the subject of those activities is a conjoined mind and body. If, instead, only acts of the mind are properly acts of the subject, then walking might be some act of mind that is mine, plus a caused or conjoined bodily activity which is the act of some other thing. If that were the case, and I were directly aware of both the act of mind and the bodily activity I would be aware of two things only one of which is me. If I were directly aware only of the act of mind then I would be directly aware only of me – but I would not be aware of my walking.

However, if walking is a proper activity of the subject, and the subject is directly aware of it, this suggests that a possible, if anachronistic, reply is available to Gassendi. It may be, he could reply, that we cannot know with certainty that I am walking on the basis of my conscious activity of walking. But if you grant that my experience incorporates my walking, and by so doing incorporates me as that of which it is a mode, I can know, non-infallibly, on that basis that I exist.

We need, in other words, to separate out the question of whether experience can provide non-question begging grounds for existence claims, from the question of whether it delivers certainty in the face of the sceptic. If we are interested just in the first question, about whether or not the Ambulo argument gives us non-question begging grounds for our existence, then Descartes' response does not close off a positive answer. We have got a reason, through our experience of walking, to believe that we exist. Certainty is another issue. It is true that we may be wrong about whether we are actually walking – we could be dreaming – but in that case we have not, on the account of engaging in conscious acting being considered, got a conscious experience of walking. All we have is an illusion of walking, and that was never claimed to offer us grounds for a proof of our existence. And note that certainty over our acts and activities does not get much easier if we limit ourselves to activities that do not involve movements of the body – which we might think are not primary activities of the self for Descartes. Consider covert activities such as guessing or supposing. Whether or not a subject is supposing, or guessing, or even judging, might seem to imply certain dispositional features: whether they are prepared to discharge the supposition, whether they lack knowledge on the matter they are guessing about, whether they are prepared to use their judgement as a reason in a argument. Given this it seems clear that one can make sense of someone taking themselves to be supposing, guessing, judging when those features do not obtain and so, when they are in fact not supposing, guessing or judging. They only have the illusion of doing so. It is no surprise that certainty is hard to come by, but it is worth noting that it is hard to come by both in relation to covert psychological acts/activities such as supposings and guessings, and overt psychological activities such as walkings and jumpings.

We will come back to whether the activity of thinking has a special capacity to secure certainty in a way that the other activities don't. But for the moment let me sum up where we have got to. I have claimed that if in engaging in a conscious action we have a direct experience of our actions which is to understood in the way that the relationalist Moorean thinks that I have direct experience of my hand, then we have available to us a form of Cogito argument that can, in a non-question-begging way, provide grounds for knowledge of our own existence.

I have not provided arguments, and am not going to, for the claim that we do indeed have direct experience of our actions, and of us acting. I think the view that in acting we have an experience of our actions which encompasses those actions, and their agents is right – and indeed may be more plausible than the parallel view in relation to visual perception. However, my interest in this paper is only to identify the space for it, and establish the conditional conclusion that if the view were right, and if thinking is the activity of an agent, then there is a non-question begging version of the Cogito.

However, before turning to objections, there is scope to emphasise an advantage that would flow from such a view of our experience of our actions – other than providing a working version of the Ambulo, and in turn the Cogito, on the assumption that thinking is a form of acting. The view has much the same advantage that tends to advertised by the relationalist about perception in general: that it concords with our sense of being in direct contact with that which we are aware of. If we did not have direct experiences of our walkings, jumpings and so on, we would face the prospect of residing in a phenomenological bubble of action awareness with the actions themselves always something that's beyond our experience of them. It is often claimed by relationists about perception that non-relationalism leaves a subject alienated from the world by ‘a veil of perception’. If there is such an alienation, then the extent of it is hugely magnified if the separation is not just between me and the external world, but between me and every one of my activities – both covert and overt. Every action I carry out – my walking, jumping, supposing and guessing – would be somehow distinct from and beyond my experience of it. That kind of picture would be decisively set aside if one accepted the relationalist view.

6. Limits and Obstacles

(i) The nature of thinking. One thing one might say in response to the above discussion is: look, I accept the Ambulo argument. I am convinced that, if we take walking to be a genuine mode or way a subject might be behaving, then in being directly aware of the walking I am directly aware of the subject walking. And if we do that, then the Ambulo is an argument I can use to prove my own existence. Suppose I am lying in a floatation tank and start to have doubts about whether I exist – perhaps I start to worry that I am merely some kind of cognitive ether and have no real existence. All I need to do is to get out and walk. If I engage in the conscious activity of walking I will have all the grounds I need to prove that I exist.

But that, the objection runs, is not going to resurrect the Cogito. It is not going to resurrect the Cogito because we've got no reason to think that thinking is an activity of a subject, awareness of which provides awareness of the subject. If I am a human being and my walking is way a human being is behaving then it is plausible to think one's consciousness of my walking is consciousness of me. But, it might be urged, thinking is something quite different. Thinking is not a way a whole human being behaves in the way that walking is – and when I am aware of my thinking I am not aware of the human being – I am aware only of the thinking disconnected from it being my thinking. It is further step, requiring rational support to take awareness of my thinking an evidence of my existence.

I have said that I am not going to argue for the view that thinking is an activity of a subject and has no reality without being a mode of the subject. And I am not. Nor am I going to argue that we should think of ourselves as human beings for whom thinking and walking are active modes in similar ways. My argument is conditional – if that is the right view of thinking and walking then the argument works.

However, I do want to urge in reply to the objection presented, that on the most common sense picture of what we in fact are, the natural thing to think about thinking is that, just like walking, and jumping, it is an activity of a human being. If we ask the question ‘What do human beings do?’ we might very well answer along the following lines: ‘Well, we walk, jump, dance, talk, think, question, argue, skip and a whole lot of other things’. Thinking, questioning, doubting, all fall very naturally into a set of activities that are given as an answer to the question ‘What do we human beings do?’

That this is a natural and common sense thing to say about thinking – that is just another on of the many kinds of activities that human being get up to – can be brought out if we look at how we qualify attributions of thinking to people. We use much the same adverbs to qualify ongoing cognitive activities as ongoing overt physical ones. We think slowly, we can get interrupted thinking, we can think frenetically and anxiously. That there is such a continuity between the overt physical activities of a subject and thinking is brought out by the fact that one of the ways you can think, is by talking. Talking is very often a way of thinking. Sometimes we think by talking to others: we often do not think the thought and then work out how to communicate it. We just talk. Sometimes we think by talking out loud to ourselves, and sometimes by talking to ourselves in what Ryle called ‘silent soliloquy’. Similarly, we can think by writing – to ourselves or to others. Of course, sometimes we think without talking or writing at all – even to ourselves. We have, however, no reason to hold that in such thinking a subject is operating in a fundamentally different mode from when she is talking or writing. Obviously, if you think out loud you need to move your mouth, or if you are working out your thoughts on paper you have to move your hands, whereas if you are engaged in silent soliloquy, or thinking without talking or writing at all, you need only engage a more restricted part of your body. It would only be if our awareness of our talking and writing amounted only to our awareness of the movement of our arms and lips that this would give us a reason to hold that there is a radical asymmetry between being aware of ourselves engaged in talking and writing, and awareness of ourselves in covert thinking. Our awareness of our talking and writing does not amount only to our awareness of the movement of our arms and lips – if it involves it at all.

We have no obvious reason to hold that there is an asymmetry between our awareness of different kinds of thinking such that when we are consciously engaged in talking out loud we are aware of ourselves in virtue of consciously engaging in such talking, but when we are engaged in covert thinking we are not.

The other thing to wonder when one worries about whether thinking is really an activity of a subject is to ask ‘what is the alternative picture?’ There is a way of talking about conscious thought that makes it sound like a kind of phenomenological glitter. On this picture there could be phenomenological stuff going on in all sorts of unlikely places – conscious ripples disturbing murky puddles on Alpha Centuri. Or perhaps the idea is that that there could be brain fragments that could carry on the activity of thinking without there being any subject doing the thinking? But that is a very peculiar view – maybe I can survive if enough of my brain does – but if all we are left with is fragments we have little reason to suppose we are left with any thinking either. Ryle talks about the ‘elastictities of uses of “I” and “me”’. He asks us to:

I agree with Ryle: the whole of me does these things, not bits of me – even if the whole of me can shrink to something quite small. Further, and similarly, it seems to me confusing to talk of ‘my legs walking’ or ‘my lips talking’. Of course, these issues will not be settled until we settle what thinking is, and settle how we can coherently talk about thinking. It may be that the conditions on thinking can be met other than by whole subjects in certain conditions. But I think it very likely that they will not be.

(ii) Knowing a thinker exists vs knowing I exist. Footnote ¹⁴ There is a second objection that might be raised even if it is agreed that there way of construing the relation between awareness of activities, and their agents, that means that the standard Lichtenbergian objection would not get any traction. The fact that you will not get ‘thinking going on’ unless you have a subject thinking – along with the fact that this is known ‘by the natural light’ – will get you knowledge that the subject exists. Still, the objection runs, you will not get anything as strong as the conclusion that ‘I exist’ – the most you will get is the conclusion ‘Someone exists’. The fact that the subject that exists is me is additional to knowing that a subject exist.

This suggests the possibility of a non first-personal, existential version of the Ambulo:

My response to this objection is threefold.

First, even if this objection is right about the inaccessibility of the first person judgement it is wrong to think that we can get only an existential conclusion. We can also reach a demonstrative conclusion – ‘this subject’ who walking I am aware of exists; ‘this subject’ whose thinking I am aware of exists:

Second, the fact that we have identified an existential and demonstrative version of the Ambulo is itself of note. If it is granted that awareness of an activity of a self is sufficient to ground awareness of a subject and warranted judgements about that subject – whether or not I know it that that self is me – we have re-positioned the gap that was supposed to be surpassed. It is no longer a gap between an act of thinking and its subject, but between a thinking subject and identifying who that subject is.

Three, we do have forms of direct awareness of activities of subject that themselves may warrant only the demonstrative and existential conclusions – such as when we see someone walk – whether we see another, or see ourselves reflected in a mirror. However, when we ourselves engage in the conscious activities of walking or thinking, when we are the walker and the thinker, our awareness of what we are is through a distinct form of awareness – an awareness we have through being the agent of the activity. If that is right then, without positive reasons to think that I cannot be walking or thinking – or that I am subject to an illusion of agents awareness, my conscious walking or thinking will always provide a warrant for judging ‘I am walking’ or ‘I am thinking’.

(iii) Hume's Intuition: Something that might still worry us about the above way of trying to secure the epistemological respectability of the Cogito is the thought ‘what happened to Hume's intuition?’ Wasn't Hume right to observe that when we introspect we find ourselves missing in some way? As Hume famously put it says:

We have, I think, all been a bit mesmerized by this quotation. There are two things peculiar about it. One, is the idea in order to catch myself I need to catch myself without my perceptions, and need to observe anything other than the perceptions. If I manifested myself to introspection in perceptions, then the way to catch myself is to catch the perceptions. Imagine Hume had been worrying about whether external objects show up in vision. Are they not somehow missing? After all, he might have argued:

Now it maybe that Hume himself would in fact say that – but this way of seeing things does not capture a common sense way to report our phenomenology of the world and our capacities to be acquainted with ordinary objects. And I don't think his quotation about the self should be reported, as it so often is, as the natural and common sense way to report our phenomenology of the self.

The second peculiar thing to note is the list of things Hume thinks we should pay attention to in our efforts to try to find ourselves in introspection. Suppose he had instead ‘entered most intimately into himself’ and stumbled across his thinking, his looking, his seeing, his calculating, his talking, his walking, his dancing and jumping, but declared that he never caught himself without any of these things, then I think we would want to reply: ‘Well, you've been there all the time; you have already stumbled across yourself.’ What Hume seems to be asking for is observation of the self bare of all it's activities; we should no more think we can experience a self bare of its activities than we should think we can experience an object bare of its qualities. If activities are ways I may be, nothing justifies the expectation that to be aware of myself I need stumble on my ‘self’ on its own, bare and stripped of its activities? Whatever the self is we are aware of it in its activities.

(iv) Certainty. It seems to me that we have, given certain assumptions about the nature of activities and what we experience when we experience them, good reason to take ourselves to have available a working version of the Ambulo and the Cogito. I will end by asking whether we have a reason, as Descartes thought, to prefer the latter to the former because it gives us certainty. Do we get more with the Cogito than the Ambulo? The Ambulo gives us warrant for existence of subjects, but will not survive doubt about whether I am really walking or just suffering an illusion to that effect. Might the Cogito do better, and so doing give us not only warrant for our own existence but certainty about our own existence?

Well that depends on whether, on the picture being presented, I could coherently be wrong about, or doubt, whether I am thinking? Suppose it seems to me that I am thinking – could I be wrong? Well, if its seeming to me that I am thinking involves my thinking ‘I am thinking’, then I cannot be wrong – I am thinking ‘I am thinking’ and so what I am thinking is self-verifying. However, perhaps there is another way to understand what would have to be going on for it to seem to me that I am thinking. Perhaps all that need be involved is an occurrence that has a feel just like this, this thinking now going on, but which is an occurrence that is not a thinking: it is an occurrence which fails a condition on thinking for one reason or another. If it is possible for there to be an occurrence that has a feel just like the feel of running through the Cogito argument, but it is not a running of the Cogito argument because it does not involve thinking at all, then a thinking subject running through the Cogito may coherently wonder whether something like is going on rather than that she is thinking through the Cogito. However, if she does so she can comfort herself with the thought that were that to be the case she is not wondering anything – wondering takes thinking; and she is not running through the Cogito – running through the Cogito takes thinking. She cannot in fact have even have engaged in the first step of the Cogito; she cannot have engaged in an act of conscious thinking. If all that is going on is a conscious non-thinking occurrence that feels like this, then she will not get her conclusion, but she will not her premises or her argument either. She will have done nothing. However, I have argued that, as long as she started with a conscious thinking – although doing so might come with meeting significant conditions – she may be able to get her conclusion without begging the question.