Sunday, 19 February 2017

Word of the Week: Realism (Philosophy)



Philosophical realism is the theory that there is a material world of 'ordinary objects' for us to perceive. In other words, realists think that there is some objective, physical reality out there - as opposed to just ideas in our minds. Realists, therefore, do not think that we are in the matrix, or mere brains in vats. However, realism has come under attack from various philosophers: some deny the existence of the material world completely (e.g., George Berkeley) or at least deny the possibility of a proof of the material world’s existence (e.g., Immanuel Kant). Perhaps more significantly, realism has come under attack from politicians: Donald Trump and his minions are intent on denying the validity of our senses. For instance, Kellyanne Conway (Counsellor to the President) recently made up a terrorist attack - the 'Bowling Green Massacre' (hint: it's fantasy) - despite the fact that such a terrorist attack was never seen or heard to take place. This situation, of resorting to mere ‘idealism’ where the world is just what we wish it to be, is what ‘realism’ tries so hard to avoid. Realists do this by affirming the possibility of veridical (truthful) experience of the material world. The likes of Plato and Trump would have us believe that both veridical and non-veridical are insignificant, since abstract 'ideas' are far more interesting, reliable and entertaining (well, at least for Trump...). But realists have set out at solving this Trumpist dilemma.

Nevertheless, realism is not without challenges, and has descended into a conflict between representative realists who assert that our perceptions indirectly represent reality and naïve realists who assert that we can be directly aware of reality. I argue that a realist theory of perception can only be properly upheld in a representative way. However, even from that standpoint, representative realists must produce a justification of why our perceptions represent reality at all. If such an argument is not proposed, then representative realism decays into mere idealism. In that vein, I will start with evaluating the aptly named theory of naïve realism before progressing onto representative realism in the second part of my critical analysis. This should facilitate a judgement as to whether it is indeed impossible to uphold a realist theory of perception.

Naïve realism is built upon the foundations that common sense provides, as Bertrand Russell (1912) and Crane and French (2017) note. This common-sense view itself is a realist theory of perception. Therefore, we can begin our analysis of naïve realism with an analysis of its foundations in common sense. The common-sense view of perception tends to assert two axioms, one of which concerns Openness and the other of which concerns Awareness. As John McDowell notes, the first axiom means that we have ‘Openness to the world’, which comes in three ways: firstly, mind-independence, or the idea that, when we perceive objects, we are perceiving objects that are independent of our minds; secondly, presence, or responsiveness to the way things are now, such that a perception of a table means that we are responding to a real, present, material table; thirdly, transparency, or perception only of mind-independent objects and absolutely nothing else. The doctrine of Awareness is defined in roughly the same way as Openness, with the caveat that humans are in actual perceptual contact with objects. For instance, according to Awareness, if we say that we perceive a table, then a common sense realist would say that we are directly aware of the object, with no need for any intermediary sense-data or representations to provide ‘stepping stones’ for us.

Martin Heidegger (1977 [1950]) noted the merits of this approach by acknowledging that we never ‘originally and really perceive a throng of sensations, e.g. tones and noises, in the appearance of things’; ‘rather, we hear the storm whistling in the chimney, we hear the three-engine aeroplane’. Therefore, when we hear ‘the door slam in the house’, we are hearing the actual door slamming and ‘never hear acoustic sensations or mere sounds’. This has the advantage that we are not acquainted with mere sensations of mental sense-data, but instead are acquainted with, to use a Kantian phrase, ‘things-in-themselves’. Thus common-sense realism allows us to find truths about reality itself, and not just our subjective perceptions. (Hence the materialist maxim of phenomenologists such as Heidegger: ‘To the things themselves!’)

However, whilst Openness and Awareness are both ‘common sense’, there are a number of philosophical problems with both axioms. In particular, the third element of Openness—namely, that we perceive objects with total and utter ‘transparency’—runs contrary to the evidence. To say that we perceive with transparency implies that any variation in our sense-experience can be explained with reference to some variation in the actual objects themselves. For instance, if a white light were shone on a green object from a certain angle, then a common-sense realist would have no choice but to say that the green object has actually turned white. This is obviously not the case – the object is still green from whatever perspective of the observer and whatever lighting is applied to the object.

Nevertheless, academic naïve realists, rather than common-sense naïve realists, solve this problem by admitting three ways in which Openness does not always equate to total ‘transparency’. Firstly, as Heather Logue (2012) argued, variations in the perceiver, such as whether the perceiver is wearing pink-tinted glasses, can lead to variations in one’s perceptions which are not present in things-in-themselves. Secondly, as John Campbell (2009) and Bill Brewer (2011) noted, there are certain conditions of perception, such as angle of the object relative to the perceiver, that can generate illusions, such as the anamorphic skull in Holbein’s The Ambassadors. Thirdly, as M.G.F. Martin (1998) and Matthew Soteriou (2000) noted, the way in which we view an object can generate illusions in our perceptions, such as looking at a cup of tea from an ‘upside-down’ perspective, generating the illusion that the liquid in the tea is suspended in the air. Although these responses are, in my view, essentially the same (as they argue that changes in the relationship between objects and perceiver can impact on the perception of the objects) they succeed in saving academic naïve realism from some of the more obviously naïve claims of its common-sense counterpart.

Yet academic naïve realism is dealt a fatal blow by appliers of the arguments ‘from illusion’ and ‘from hallucination’, such as Howard Robinson, C.D. Broad, J.J. Valberg and P.F. Snowdon. Moreover, representative realists such as Howard Robinson have melded the two arguments against naïve realism to also argue for representative realism. The argument progresses as follows:
(A) Base Case: (a) illusions and hallucinations are not really cases of awareness of ordinary objects; but rather (b) are cases of awareness of non-ordinary (i.e. non-material) sense-data.
(B) Spreading Step: veridical experiences, illusions and hallucinations are all to be given the same account.
(C) Conclusions: (c) no perceptual experience is the case of awareness of an ordinary object; bur rather (d) all perceptual experiences are the cases of awareness of non-ordinary sense-data.
From the apparently obvious premise (A), representative realists have demonstrated that we never have direct awareness of ordinary objects and instead always have direct awareness of non-ordinary sense-data. Sense-data, for our purposes, are (1) objects of direct awareness, (2) bearers of sensible qualities (i.e. that we can ‘sense’ or ‘perceive’), and (3) objects of a radically different ontological kind from ordinary objects (i.e. rather, they are non-ordinary objects). Sense-data, therefore, are the immaterial, mental representations of the material, non-mental reality beyond.

However, a naïve realist, who would prefer us to directly perceive reality rather than indirectly perceive it via sense-data, could attack the Base Case of the argument. Firstly, a naïve realist would argue against the notion in (A)(a) that ‘illusions are not really cases of awareness of ordinary objects’. (Note that the naïve realist is ignoring the argument ‘from hallucination’ for now and is concentrating on the argument from illusion’). Naïve realist Bill Brewer (2011) argues that an object, under certain material conditions, will appear to us in an illusory way. For instance, the placement of a pencil at an angle into a glass of water may make the pencil appear bent. For Brewer, we are still aware of the pencil, but the pencil is placed in such a way that it appears as if it is bent. It is not sense-data that tell us this, but the physics of refraction that tell us this. Therefore, Brewer contends, we are still directly aware of the ordinary object, albeit in an illusory way. On this basis, a naïve realist contends, the Base Case collapses and the argument against naïve realism and for representative realism falls apart. Nevertheless, even if illusions are still cases of direct awareness of ordinary objects, the argument contra naïve realism might still work on the basis of hallucinations: all that is needed is one base case of non-awareness of ordinary objects, whether that involves illusions or hallucinations, for the argument to suggest that we never have awareness of ordinary objects. Therefore, for naïve realist to uphold their case, they have to also take down the part of the argument that deals with hallucinations, or totally non-veridical experiences, such as dreams.

J.M. Hinton (1973) and John McDowell (1982, 1987) try to take down the Spreading Step’s notion that ‘veridical experiences […] and hallucinations are all to be given the same account’. To do this, they propose the disjunctive theory, which suggests that there is something intrinsically different about veridical experiences from hallucinations. Veridical experiences are cases of direct awareness of ordinary objects whereas illusory experiences are not cases of direct awareness of ordinary objects. (Of course, therefore, naïve realists have to accept the Base Case on this one, so they are just attacking the Spreading Step.) Therefore, for naïve realists, we cannot accept the Spreading Step’s principle of giving ‘the same account’ to both illusory and veridical experiences, since they are intrinsically different, or ‘disjunctive’. However, veridical and illusory experiences are often subjectively indistinguishable. For instance, if one is hallucinating that one is looking at a polar bear in the Arctic, then this hallucination may be as vivid as a genuine perception of a polar bear in the Arctic. Or, at the very least, it is very difficult to tell when one is dreaming and when one is not. Consequently, it seems that the argument from disjunctivism falls on its basic assertion that there is something that obviously distinguishes hallucinatory from veridical experiences for the perceiver. As Delia Graff argues, hallucinations feel, for the perceiver, almost identical to veridical experience, such that the Spreading Step is correct in assuming ‘that veridical experiences, illusions and hallucinations are all to be given the same account.’

Therefore, it seems that representative realists have the theory with the greatest explanatory power: when we are having hallucinatory experiences, representative realism teaches that the reason why the experience is so vivid is that we are really and directly acquainted with objects. These objects, however, are just sense-data, which may or may not correspond with the real world. A.J. Ayer, for instance, contended that all experiences can be explained in terms of sense-data. However, sense-datum theorists such as Ayer have admitted that sometimes this would involve (using Robert Nozick’s terminology) ‘subjective conditionals’. For instance, a subjective conditional might involve the claim that ‘if I were standing high on the right side of Hans Holbein’s The Ambassadors, then I would see an accurate painting of the human skull’ but ‘if I were looking front-on at Hans Holbein’s The Ambassadors, then I would see a strange, grey, distorted shape’. In the case of this painting, subjective conditionals must be employed to explain why the skull is anamorphic, i.e. it appears in different shapes from different perspectives. However, this suggests that the conditionals must always hold, which is implausible if sense-data are just mental cognitions. If sense-data are mental, then our perceptions should be chaotic and dream-like, without a tangible pattern. This requires a representative realist to posit the existence of the material world to be perceived, such that veridical experiences are still veridical and illusory experiences are still illusory. However, as A.J. Ayer notes, there is no a priori proof for the existence of material ordinary objects. Nevertheless, we can assume that such objects exist, as they make uniform perception possible. In other words, we can pragmatically but not logically demonstrate the existence of material objects. Without such an assumption, science would have to assume that objects, when unperceived, might not really exist; with such an assumption, we can happily leave our homes without worrying that the homes will vanish when we are not perceiving them. In a way, the existence of matter is, to use Kantian terminology, ‘empirically real’ for all scientific purposes, but, for want of a logical proof, ‘transcendentally ideal’.

On balance, representative realism deals naïve realism a fatal blow with the arguments from illusion and hallucination. Even though illusions can be explained by naïve realists as examples of direct awareness of ordinary objects under certain material conditions, hallucinations are clearly non-veridical and, at the same time, subjectively indistinguishable from veridical experiences. Representative realism solves this dilemma by explaining all experience in terms of sense-data and not just ordinary objects. Representative realism, however, similarly lacks any a priori demonstration of its veracity. Nevertheless, there certainly are a multitude of a posteriori indications of the plausibility of representative realism, as its axioms are presupposed in much scientific activity. Pragmatically, realism is absolutely necessary lest we admit that objects may magically vanish if they are not being perceived. Realism may not be a perfect theory, both in its naïve and not-so-naïve forms, but it certainly is not impossible to uphold a realist theory of perception. So the Trumpists may have lost - the real world does exist, perception can give us veridical experiences, and truth matters.

Friday, 10 February 2017

Word of the Week: Compatibilism

 Portrait of compatibilist Immanuel Kant

Note: key words or phrases coloured in orange constitute the unfettered determinism that Kant is opposing, whilst everything coloured in purple constitutes the theory that Kant is advocating - namely, compatibilism. 

Last week, I asked whether determinism, or the theory that humans have no free will, is convincing. This week, I consider whether compatibilism, or the view that free will and the rigid laws of nature are compatible, is reasonable. There are a range of compatibilist thinkers out there, but today I'll focus on one of the the most convincing compatibilists - Immanuel Kant. In his Critique of Pure Reason, his Grounding for the Metaphysics of Morals, and his Critique of Practical Reason, Kant suggests that determinism may have an element of truth. But he also provides two arguments against unfettered determinism and for compatibilist free will. Kant thus occupies a middle ground between advocates of complete free will and hard determinists. 

Firstly, Kant argues that the ‘autonomy of the will’ is the ‘supreme principle of morality’. Determinists would like to argue that the will is heteronomous, or that the will, when making judgements, always looks to the ‘objects of volition’, or desires such as egoism and concern about one’s reputation. However, Kant argues that the will is actually autonomous: when we say ‘I ought not to lie’, even the most committed villain would not assert that this ‘ought’ entails the condition that ‘I would maintain my reputation’. Therefore, any imperative, Kant asserts, is not hypothetical, but is rather categorical. In more common language: an imperative is absolute and not relative to the situation. So the will should not look to sentiments in reaction to the external world, but rather to the reason given to us by our internal mind. As our moral will is created internally rather than externally, and is autonomous rather than heteronomous, Kant insists that the will is free. John Locke and David Hume used similar arguments in their emphasis on internal rather than external causes as the key determinants of the will. For compatibilists, therefore, our will is both free and determined: we lack ‘external impediments’, and in that sense are free, but still have to listen to the internal maxim of the categorical imperative. Here is my formal presentation of Kant's argument, using my own logic:
  1. If determinism rules out free will, then the will is not free;
  2. If the will is not free, then it is heteronomous;
  3. If the will is heteronomous, then it is not autonomous;
  4. If the will is not autonomous, then moral imperatives are hypothetical (relative);
  5. If moral imperatives are hypothetical, then they are not categorical (absolute);
  6. But moral imperatives are categorical, e.g. 'I ought not to lie';
  7. Therefore, moral imperatives are not hypothetical;
  8. Therefore, the will is autonomous;
  9. Therefore, the will is free, i.e. free will exists;
  10. Therefore, determinism does not rule out free will.
However, this suggests that the will is always categorical, when in fact people are perfectly capable of willing something with reference to a multitude of external, sentimental causes. Examples of non-moral (i.e. non-categorical) causes include the enhancement of one’s reputation instead of the adherence to a Kantian ‘universal law’. So internal and external causes, together with rational and sentimental causes, are not clearly distinguishable, suggesting that compatibilism may be misled. Kant relies on the autonomy of the will to justify the categorical imperative and, in turn, free will.

The second Kantian argument for free will worth noting is his ‘third antimony’ found in the Transcendental Dialectic (from Kant's Critique of Pure Reason). Kant contrasts free will and determinism by analysing the arguments for both. On the one hand, Kant acknowledges that free will gives humans control over their actions, but on the other hand Kant admits that all objects presented to our senses are predetermined, including human action and the will. However, here Kant refers back to the distinction he made in the Transcendental Aesthetic between the noumenal world (or the real world) and the phenomenal world (or the apparent world). Whilst all objects occupying space and time are part of the phenomenal world, Kant suggests that space and time are just subjective intuitions. Thus the noumenal (real) world may not really be that which we perceive when we see, touch, smell, taste or hear objects. Instead, the noumenal world lies beyond our senses, and is thus totally unknowable, since Kant agrees with the empiricists that we cannot prove, metaphysically, the existence of anything without referring to our senses. Although Kant disagreed with Newton on many issues (such as the very noumenal existence of space and time, which Kant denied and Newton affirmed), they agreed that experience is the limit of our knowledge concerning what seems to exist. So Kant said that, whilst determinism is a property of the phenomenal world, free will may be a property of the noumenal world. Here I use Leibnizian logical laws to explain Kant's point:

  1. According to Gottfried Leibniz's Indiscernibility of Identicals, if x is identical to y, then for every property F, object x has F if and only if object y has F; i.e.  x=y → ∀F(Fx ↔ Fy); 
  2. Therefore, if we let p equal the phenomenal world and n the noumenal world, then n=p implies that for every property D, or unfettered deterministic causation without a jot of free will, n has D if and only if p has D; i.e. n=p → ∀D(Dn ↔ Dp).
  3. According to Leibniz's Identity of Indiscernibles, if, for every property F, object x has F if and only if object y has F, then x=y; i.e. ∀F(Fx ↔ Fy) → x=y;
  4. But we can never describe space or time in concrete terms, but only in incomplete analogies such as 'time [is] a line progressing ad infinitum' (Critique of Pure Reason, 1781); this is an incomplete analogy because, whilst the parts of a line are simultaneous, time occurs sequentially; in other words, whilst it is true that our intuitions of space and time are properties of the phenomenal world, it is not necessarily true that intuitions of space and time are properties of the noumenal world. So, where the logical symbol '¬☐' means 'not necessarily' and the property I means 'our intuitions of space and time', it's not necessarily true that our intuitions are properties of the noumenal world; i.e. ¬☐{∀I(In ↔ Ip)};
  5. Therefore, and with reference to Leibniz's Identity of Indiscernibles, it is not necessarily true that the noumenal world is identical to the noumenal world; i.e. ¬☐{∀I(In ↔ Ip)} → ¬☐(n=p);
  6. Therefore, and with reference to Leibniz's Indiscernibility of Identicals, it is not necessarily true that deterministic causation without any free will is a property of both the noumenal and phenomenal world; in other words, merely phenomenal determinism does not preclude noumenal free will; i.e. ¬☐{∀D(Dn ↔ Dp)}.
However, this argument can be criticised using Kant’s original assertion in the Transcendental Aesthetic—namely, that the noumenal world is unknowable. As such, whilst determinism is at least a property of one of these worlds (the phenomenal world), we can never say for certain whether free will is a property of any world (the phenomenal world or the noumenal world). Determinism is, in all probability, more ‘real’ than free will, such that Kant’s compatibilist balance between free will and deterministic causation cannot, realistically, be exactly 50:50.

Overall, whether you accept Kantian compatibilism rests on whether you accept the possibility of categorical imperatives, as well as the possibility of any speculation at all regarding the noumenal world. The choice is yours... or is it?

Disclaimer: Kant is notorious for the myriad of possible ways of viewing his work. So this post gives my personal interpretation of Kant's work, which I leave open to debate amongst Kant scholars with different interpretations and other students of the revolutionary Königsberg philosopher. This is, after all, what Kant said 'enlightenment' was all about.


Wednesday, 1 February 2017

Word of the Week: Determinism

Determinists try to argue not only that the future is ‘determined’, but also that we have no free will over this matter. This has driven many supporters of free will to despair, as they can’t get to grips with the justification of determinism. So I’m going to see whether there’s any reason to believe that we have no free will.

In the deterministic view, every human action is not only influenced by what has come before, but is rather predetermined. Let P be a true description of the complete state of the universe at some point in the past; let L signify the laws of nature; and let F be any given event in future. According to determinism, P and L predetermine F. In other words: if we had total knowledge of the state of the universe 4 billion years ago (which is P), we could deduce, from the laws of physics (which is L), what I would eat for breakfast tomorrow (which is F). This is the thesis of determinism: due to the uniformity of the laws of nature, every single event, including every single human action, is predetermined. The following argument is my adaptation of Carl Ginet and Peter van Inwagen’s consequence argument, as articulated by Kevin Timpe. The argument begins by identifying two axiomatic modal operators, alpha and beta:
Axiom 1 (Alpha): ☐p > Np
Axiom 2 (Beta): {Np & N(p > q)} > Nq
The  ‘☐’ sign can be read as ‘It is logically necessary that…’ whilst ‘Np’ means ‘p is true and no one has, or ever had, any choice about whether p is true’.  Necessity, for our purposes, means ‘true in all possible worlds’, i.e. it cannot not be true. The symbol N, in turn, means that our free will has no say in the matter. The sign '>' means 'implies that'. Thus, the Alpha operator says 'it is logically necessary that p', which implies that we have no choice over necessary things. The Beta operator looks complicated, but it basically means that, from the premises that we have no free will over p and over p implying q, we can deduce that we have no choice over q. The argument then progresses to spell out the ‘definition’ (meaning) of determinism, its ‘corollary’ (synonym) and ‘axioms’ (assumptions or premises).
Definition 1: ☐{(P & L) > F }
Corollary 1: ☐{P > (L > F)}
Axiom 3: NP
Axiom 4: NL
So Definition 1 says that it is necessary that, given knowledge of a past situation P and the laws of nature L, it is possible to logically deduce any event F in the future. So this says that determinism is true – but it doesn’t say that free will is false! So the argument continues… Corollary 1 just rearranges Definition 1. Axioms 3 and 4 say that we have no free will over the past and the laws of nature. It follows, from Alpha and Corollary 1, that Proposition 1 is true:
Proposition 1: N{P > (L > F)}
Proposition 1, you may notice, looks just like Corollary 1. But it replaces the ‘☐’ sign for a ‘N’ sign. It’s just saying that, as well the past necessarily implying that the laws of nature cause the future, we also have no free will about this causation. It then also follows, from Proposition 1, Axiom 3 and Beta that Proposition 2, as follows, is true:
Proposition 2: N(L > F)
Now that we’re progressing onto the really important propositions, I’ll just say exactly what the symbols mean. Proposition 2 reads as ‘The proposition that the laws of nature entail any given future event and no one has, or ever had, any choice about whether the proposition is true’. It then also follows, from Axiom 4, Proposition 2 and Beta, that Proposition 3, as follows, is true:
            Proposition 3: NF
Proposition 3 reads as ‘F is true and no one has, or ever had, any choice about whether F is true’. In other words, any event F in the future has nothing to do with free will. We have absolutely no control over the truth of event F, and are ourselves determined by past circumstances P and laws of nature L. So this argument lends support to the claim that humans have no free will.

Nevertheless, the most devastating attack on Beta perhaps comes from Andrew M. Bailey, who says that Beta is not self-evident in any way. Bailey allowed p to be ‘Hurricane Katrina ravages the gulf coast’ and q be ‘Countless lives are lost’. According to Beta, based on Np and N(p > q), we are meant to infer that Nq. This means that, if we are not responsible for the hurricane, then we are not responsible for the loss of life. But this, as Bailey rightly points out, does not conform to our normal ethical standards of taking responsibility for saving lives in spite of natural disasters which are out of our control. We may not have free will over the hurricane, but Beta is wrong in inferring that we do not have free will over how we react to the hurricane. Therefore, Beta is not true in all cases, so should not guarantee the truth of the deterministic claim NF either. Nevertheless, this critique relies on there being some moral responsibility of saving lives, not necessarily free will. Free will and moral responsibility are not necessarily tied to one another, so even if we have responsibility to save lives due to a hurricane, it is entirely possible that we will be unable to save those lives for one reason or another. We may want to respond to situations, but that does not entail our having any control over such situations. Therefore, Beta is, albeit a controversial, a nevertheless strong operator, which is surprisingly resilient to attacks. Yet its controversy is enough to make us ask if free will might, in actual fact, exist, giving rise to the free will alternative to determinism.


Next week, I’ll consider those who advance the ‘free will’ thesis, to see whether they are as strong as determinism’s powerful logic. Advocates of this view are often called ‘compatibilists’, as they think that free will and determinism are compatible – a view that your average determinist will vehemently deny.