Real Appearance

 

Philosophers are apt to speak of appearance and reality as if the two are opposed. But this is too simple: appearances are necessarily real and realities necessarily appear. The dichotomy no doubt has its origin in cases of perceptual illusion—the stick in water appears bent but in reality is not.  Here is the appearance and there is the reality, the two running on separate tracks. Then the philosopher converts this distinction into a general thesis about the world, dividing it into Reality and Appearance (capitalized). Reality may be said to be the totality of appearances (idealism) or to transcend appearances (realism). The concepts are clear, and clearly distinct, now we just have to make use of them to formulate grand theories. But let’s look at them more closely, in particular, their alleged mutual exclusiveness. Is this really a good way to carve things up? How exactly are the two concepts related?

First, appearances are real: when it seems to you that the stick is bent it really does appear bent. The stick is one thing–an element of reality–and the appearance of it is another–a distinct element of reality. Both are real, and equally real, but they are not the same. We might say that the appearance is not part of the reality of the stick, but neither is the stick part of the reality of the appearance (it could exist without the stick). The appearance is part of the real world of appearances: impressions, presentations, ways things seem.[1] We might say it is part of the world of psychology—the way things strike minds. So why are we opposing appearance and reality when appearances are constituents of reality? They are just not parts of that reality—the part that the stick exists in. They are not parts of physical reality, as we like to say, though they are parts of reality as a whole. Moreover, they are parts of objective reality: they objectively exist just like other aspects of the mind (the subjective is also objective in this sense). They are objective facts: it really did objectively seem to me that the stick was bent. Appearances exist at certain times and places, have causal powers, and can be measured—they are not phantasms or fictions.  If we say, for example, that colors belong to the realm of appearance, we in no way assert that they are not objectively real—they are just not included in the part of reality that (so-called) physical objects occupy. Being assigned to the realm of appearance is not a form of ontological demotion.

But now things are about to get murkier: do appearances have a reality that transcends their appearance, and do realities necessarily appear? I think the answer to both questions is yes, though the matter is hardly straightforward. Taking the second question first, is it true that every real thing has an appearance? Not a perceptual appearance, to be sure, because of numbers, moral values, thoughts, and appearances themselves. We don’t have visual impressions of these things. Still, they are presented to our minds in a certain way: we have (as Frege would say) modes of presentation of them. They don’t appear to us naked, just as they are in themselves, with no infusion from our minds; they appear to us in ways conditioned by our modes of sensibility and intelligibility. We grasp numbers, say, in the way human beings grasp them—as finite fallible embodied creatures. Not all conceivable beings grasp them this way: some may survey them far more extensively and intuitively. Animals may grasp numbers in a way different from the human way—more primitively, we suppose—and superior intelligences may grasp them differently from us (and we vary among ourselves according to our mathematical prowess).[2] Appearances are not always perceptual, yet they exist nonetheless. Similarly, unobservable entities like atoms have an appearance—how we represent them in thought. Maybe we use images to do so or maybe discursive symbols, but we grasp them somehow—under some mode of presentation. Conceivably, other beings might grasp them differently, according to their cognitive structure.[3] But what about unknown realities—do they have appearances? What about reality before minds came to exist? Here I think we can say that such unknown objects have conditional appearances: there is a way they would appear to minds of such and such a type. This is a determinate fact about them: every reality has a way (a set of ways) of possibly appearing to cognitive beings. The reality constrains the appearance, without exhausting it, and the appearance belongs to the reality as an objective fact. There is a fact of the matter about how unknown particles would appear to the human mind—or numbers not yet contemplated. There is no such thing as appearance-free reality: real things necessarily have appearances built into them (sometimes conditionally). The same is not true of fictions: they don’t always have a determinate way they would appear, because they don’t always have a determinate nature. How would Sherlock Holmes appear? But things with a determinate objective nature thereby have a determinate mode of appearance, which is relative to the being appeared to. Appearance is thus woven into reality not imposed on it arbitrarily from outside. There is no appearance-neutral reality.

As to the first question, I believe that appearances always transcend their appearance. There is always more to them than there seems. This is because they are themselves realities. For example, the stick seeming bent to me at time t is an experience with a nature—a visual experience occurring at a certain time with certain causes and effects. Not everything about it is apparent to my mind. Maybe we can say that it appeared to me a certain way—the appearance did—but it wasn’t exhausted by this second-order appearance. It is not easy to argue for this view, but I think it has intuitive plausibility: why should the part of reality consisting of appearances be different from the rest of reality? Appearances are real too, so they should have a nature that transcends their appearance—this is part of realism about appearances. Appearances are natural occurrences or states with a basis in the brain, a history, and a science proper to them: why should they lack a nature that goes beyond how they naively seem? An idealist about appearances would suppose that they are nothing but how they appear, but a realist will insist that nothing real reduces to its mode of presentation—even modes of presentation themselves. Senses, to use Frege’s terminology, have a nature of their own not exhausted by how they strike the mind (and the same is true of these events of striking the mind). In other words, real things are never just a matter of how they strike a mind, which includes appearances.

Let me put it this way: objective reality always has subjective appearance built into it, and subjective appearance always has objective reality built into it. The two are intertwined not separate orders of being. This is not to say that appearance and reality collapse into each other, or that the concepts are not genuinely distinct. The matter is subtler than that—harder to comprehend. Let’s look at logical form. The concept of appearance is a three-place relation: x is an appearance of y to z. An appearance is always an appearance of something (cf. Brentano)—in the simplest case a perceptual appearance is an appearance of a physical object (or a merely intentional object in the case of hallucination). But it is also true that every appearance is an appearance to someone—a sentient being at a minimum. The word “appear” always goes along grammatically with the prepositions “of” and “to” (or some equivalent): that is its logical grammar, as Wittgenstein would say. But no such thing is true of the word “real”: when something is real it is not real of something or to someone. Realities are not relations between themselves and something else that they are of, and they are not inherently relations to conscious beings. So the two concepts are vastly different; they correspond to quite different aspects of the world. Yet they are necessarily joined at the hip, since everything real has an appearance (even if only conditionally) and every appearance has an underlying reality (if we are appearance realists). You can’t be an idealist who denies realities beyond appearances, and you can’t be a realist who denies the reality of appearances. For appearances are realities with an objective nature and realities are things that necessarily appear in a certain way. Reality is woven into appearance and appearance is woven into reality. To put it simply, appearances are real and realities appear. There is no possible conception of the world that denies these truths: we cannot conceive of reality without a determinate mode of appearance, and we cannot conceive of appearances without an underlying reality. There is no appearance-independent conception of reality, and there is no reality-independent conception of appearance. This basic truth stems from the recognition that appearances are not opposed to reality but part of it, and reality is not something that floats free of appearance like the Invisible Man (a Kantian point). We can thus understand why some philosophers maintain that reality is appearance and others maintain that appearances are somehow illusory or fictitious (there are really only brain states). But reality is always more than appearance, while never departing altogether from appearance. Idealism (in one form) is not true, necessarily so, but a realism that seeks to detach the world from all appearances is likewise doomed. The phenomenal has its noumenal aspect and the noumenal has its own (counterfactual) phenomenal aspect. Reality and appearance are indissolubly connected, though logically and metaphysically quite distinct. We would do well to reform our metaphysical language to speak of appearance-reality and reality-appearance: unities not dualities. Everything that exists is a hybrid of the two, or better a fusion.[4]

 

[1] If the world is everything that is the case, then appearances are part of what is the case—one type of fact among others. It is not that appearances lie outside the world; indeed, some philosophers have thought that appearances are the only thing that exists in the world. In any case, they are of the world—part of the totality of facts.

[2] We can thus say that every reality corresponds to many possible appearances, because appearance is a relation between an objective entity and a subjective viewpoint (this may be entirely intellectual). We might even agree that every finite reality has infinitely many possible appearances. Still the reality constrains the appearance, as reference constrains sense.

[3] This has a bearing on what is called the “absolute conception”, the way we grasp the subject matter of physics. This should not be taken to be completely appearance-neutral, since it depends on our specific cognitive and linguistic make-up, though not perhaps on our sensory make-up. Particles have their creature-relative modes of presentation.

[4] This is one of those cases in which trying to state the metaphysical truth taxes our language, no doubt because it was forged in conditions of metaphysical confusion.

 

 

Replicators

 

 

The concept of replication is central to biology. For instance, Richard Dawkins makes heavy and illuminating use of it in The Selfish Gene and elsewhere. Here I will offer some analytical remarks about the concept, with special reference to the question of what counts as replication: what is replication and how far does it extend? Let’s start with word “replica”: the OED gives us this, “an exact copy or model especially on a smaller scale than the original”. One thinks of model planes or effigies of famous people. Two elements stand out in this definition: the idea of an “exact copy” and the idea of an “original”. There must be (a) a close resemblance and (b) the replica must be derived from the original. Close resemblance alone is not sufficient to be a replica, or else random similarity would suffice; there has to be some sort of relation of dependence whereby the replica is derived from the original. Thus atoms are not replicas of each other, despite close resemblance, because there is no process of derivation leading from one to the other. The entry for “replicate” captures this nicely: “make an exact copy of; reproduce”. Replication is a process in which a copy is reproduced from an original: the original is part of the causal story about how the replica came into existence. Replication is a generative process that produces exact similarity. Finally the OED turns explicitly biological for “replicate itself”: “(of genetic material or a living organism) reproduce or give rise to a copy of itself”. Again, two elements stand out: (a) both organisms and genes are said to replicate themselves, and (b) the original “gives rise” to the copy—causes it, generates it, brings it into being. So self-replication is not the same as replication in general: some replication is driven by the original itself, while some derives from an external source, possibly a conscious agent. There is internally driven replication and externally driven replication: the former is performed by genes and organisms, while the latter requires an external agency, such as a photocopier or an artist or a modeler.[1]

One might be tempted to suppose that replication constitutes the essence of the biological: it is necessary and sufficient to qualify something as biological. The world falls into two big categories, the biological and the non-biological, and replication is the feature that makes the difference. However, there appear to be counterexamples to such a criterion of the biological. First, it is not clear that replication can’t occur outside the biological realm, as with the formation of crystals. Here we have derivative duplication but the process is entirely chemical and inanimate. Perhaps it could be argued that this is not replication proper because of some further ingredient in the notion (it is not purposive, for example), but on the face of it this looks like non-biological reproduction—exact copies produced from an original. Second, it is not clear that everything biological is replicative: isn’t sex biological, but is sex a replicator? What about homeostasis or digestion or respiration? In what sense could these be described as replicators? They are not like organisms or genes, discrete entities that generate copies of themselves; they are more like processes in which such entities participate. They aren’t spatially bounded things, chunks of matter, lumps of stuff. So it looks like replication is not the very essence of the biological, just one of its chief features. There are replicating non-biological entities and there are biological non-replicating entities.

But let’s examine this more closely. Putting aside the crystals type of case, is it really true that only organisms and genes replicate, as the dictionary suggests? What about cells? Surely they replicate, using a mixture of internal and external machinery: they literally divide into exact copies. They self-replicate according to endogenous and exogenous factors. And don’t larger bits of tissue also sometimes replicate—such as the re-grown tails of lizards? In fact not much of this sort of macro-replication goes on, but it is logically possible that it should, and the result would be bona fide replication. Body parts could be replicas of earlier body parts generated by some sort of constructive process. What about organs of the body—can we say that they too are replicators? Well, they don’t produce copies of themselves by division, like cells, but isn’t there an indirect way that they work to produce copies of themselves? That is, an organ contributes to the success of an organism in such a way that an organ just like it crops up in the next generation. It does this because the body in which it lives contains genes that construct organs like it. An efficient heart in one animal leads to a similar heart in its offspring because that heart helps the organism reproduce and pass on its genes. We can thus say that the original heart plays a causal role in the process whereby a copy of it comes to exist in a later generation. It isn’t just a coincidence that the heart of an offspring animal resembles the heart of its parents; the latter heart “gives rise” to the former heart (with a little help from the genes). The whole organism replicates itself by exploiting the genes, and the organs that make it up do much the same. So the organs are replicators too—they engage in the process of producing copies of themselves. The same holds for height, weight, musculature, eye color, etc.—all these are things that get replicated.

It is no different for mind and behavior. Mental capacities and behavioral dispositions get passed on according to their contribution to an organism’s reproductive success: excellent memory, acute vision, sound reasoning, a tendency to strike while the iron is hot, etc. In so far as these are genetically encoded (i.e. very far), they get replicated in the usual way. This is not to say that genetic transmission is the only way to replicate: there is also behavioral contagion, learning, “osmosis”, meme transfer, etc. That is, mind and behavior can be replicated by several avenues and mechanisms not just by genetic transmission, so it is wrong to limit the concept of replication to the genes and whole organisms in the manner of the OED; we need to include much more under the concept.[2]The important point is that the concept is more elastic than the dictionary allows, at least so far as its entry for “self-replication” goes. In fact, replication is a common feature of a lot of biology: organisms, genes, cells, cell clusters, organs, mental traits, and behavioral dispositions. These items vary in mechanism and type of entity, but the concept of replication is common to all of them. The concept is abstract and inclusive (rather like the concept of natural selection).

This relates to something I have argued elsewhere, namely that genes are not the end of the line when it comes to selfishness.[3] Without repeating what I have already said, the point is that the biosphere, or parts of it, acts as the beneficiary of genetic and organismic action: the biosphere persists in virtue of the genes and their host organisms, because these entities enable biological processes and traits to continue. The point I am making now is that we can extend the notion of replication to include such processes and traits. Take sex: it continues in the world by a process of replication—past sex gives rise to future sex. The future sex is just like the past sex (a replica or duplicate or copy of it), and it is the past sex that “gives rise” to the future sex. Why? Because sexual behavior leads to the survival of organisms that engage in it, along with genes that transmit it. The sexual behavior of an elephant, say, is a copy of previous elephant sexual behavior, where the earlier behavior is part of the causal story of how later sexual behavior in elephants came to be. Sex is replicated across the generations just like anatomy and physiology. And it is subject to competition and natural selection in the same way too: the better you are at it, the better you do in the reproduction races. So these biological phenomena act as replicators (aided by the genetic and organismic machinery)—they are repeated, reproduced, duplicated. Thus they are not just beneficiaries of the activities of genes and organisms; they function in a replicative manner. Sex survives because it does well in a biosphere governed by natural selection, and it is passed on by a process of replication. Homeostasis is the same: it is duplicated across the animal kingdom because it works well as an aid to survival and reproduction. So traits like sex and homeostasis manifest the attributes common to biological phenomena: they can be beneficiaries of other entities (genes and organisms), and they are subject to replication. They become increasingly numerous by a process of replication. In addition there can be variations in the traits that give natural selection something to play with: a random mutation in sexual anatomy or behavior can lead to greater reproductive success than rivals, as is also the case for homeostasis. So, to use my preferred terminology, the biosphere (or sections of it) constitutes biological bedrock—the ultimate self-replicating, selfish, naturally selected stuff of biology. Not genes and not organisms: they exist one or two levels up. In particular, the gene functions as a survival machine for the biosphere, an aid to its propagation. This is not to say that the gene is not unique and central in its own way—as indeed the organism is unique and central in its own way—but it is not the be-all and end-all, biologically speaking. If I were to coin a term for these ultimate realities, I would call them “biemes”: biemes are such relatively abstract things as sex, respiration, digestion, homeostasis, growth, locomotion, sensation, thought, and so on. These things stay in existence because genes and organisms keep them in existence, and they do so because they are useful to the survival of both genes and organisms. There are thus three pillars of biological reality: organisms, genes, and biemes—with cells and organs sandwiched in between. The organisms serve the genes and the genes serve the biemes (the biosphere in effect). The most visible things are therefore the least basic: from the conspicuous organism to the microscopic gene to the abstract bieme. All are equally real, and equally vital, but the third has a claim to being the most fundamental.[4]

 

Colin McGinn

 

 

[1] Strictly speaking, replication is a three-place relation: “x is a replica of y in respect z”. Nothing is ever a replica in every respect. In the case of genes there are three respects to consider: chemical composition, functional properties, and information content. In principle replication might preserve one of these but not the others: for example, same information but not the same chemical composition or functional profile. In practice the three go together, but conceptually they are quite distinct. We can imagine possible worlds in which genes shed their chemical features and yet preserve their functional and informational features as they get passed on. We also have the notion of partial replication as well as complete replication (itself inevitably partial).

[2] I would also apply the concept of replication to historical phenomena: sometimes a population copies another population and hence produces a common historical change, e.g. revolutions or wars. This is large-scale replicative behavioral contagion. Given this, I would be happy to rate history as a branch off biology (it is really psychology writ large).

[3] See “The Selfless Machine” and “The Selfish Biosphere”.

[4] I see no competition between this perspective and standard selfish gene theory, merely a shift of emphasis. The gene can retain its starring role, though it works for another level of reality (like shady money men lurking in the shadows of movie production). The organism benefits the gene, but the gene benefits the biosphere (including the process of natural selection itself).

 

 

Two Concepts of Freedom

 

It is hard not to feel the pull of both of the standard positions on free will. On the one hand, it seems right to say that a free action is one that is in accordance with the agent’s desires, as opposed to one that is forced on the agent in some way (the OED defines “free” as simply “not under the control or in the power of another”). This is quite compatible with determinism–physical, psychological, or divine. On the other hand, it seems right to insist that freedom requires the ability to do otherwise, which is ruled out by determinism. If an agent has no alternative to acting as he did, how can his act be free? But surely the future course of nature is always necessitated by antecedent conditions, so there are no alternative actions the agent could have performed. Thus the will is both free and not free, a contradiction. Depending upon what conception of freedom we choose to adopt, we get different answers to the question of whether free will exists. But it is assumed that there is a single thing (denoted by “free will”) over which the combatants are contending.

I want to suggest that this debate is afflicted by a methodological problem, and once this problem is fixed the solution drops out quite naturally. The problem consists in extracting the word “free” from its normal linguistic context and trying to analyze it in isolation. In fact, there are two very different notions expressed by standard locutions, which generate different answers to the question whether free will exists. Both answers are correct, so that one type of locution has application while the other does not. The locutions are “free from” and “free to”. We say that an agent is free from constraints or influences that potentially limit his range of actions: illnesses, obligations, engagements, coercion, upbringing, genes, or divine interference. In this vein we can sensibly ask if the agent is free from his desires and free from his physical condition (including his brain states): here the answer appears to be universally in the negative. I am not free from my own psychology or my own physiology—though I may be free from external coercion or prior obligations or God’s dominion. The determinist adds up all the antecedent states of the world and declares that we are not free from this totality. Again, this seems logically permissible: we simply ask whether my freedom-from extends to all of the factors bearing down on me, specifically my mental and physical states. And the answer is clear: I am not free from all of that. I don’t have that kind of freedom. In the relevant sense, I could not have acted otherwise (though there are perhaps other senses in which I could have acted otherwise, e.g. I could logically have had a different psychology). Put simply, we don’t have freedom from the past—the locution “free from” does not apply to the totality of past facts (though it applies to various subsets of these facts). It is quite true that we are free from X for many values of X, so we are free relative to these values, but we are not free from all values of X. We don’t have complete freedom-from. So we can forget having that kind of freedom. Determinism rules out freedom-from.

But it doesn’t follow that we don’t have freedom-to. I have freedom to do Y if I can act on my desire to do Y. The locution “free to” allows application in conditions in which I do as I please, as opposed to acting against my desires because of external (or internal) coercion. That is what “free to” means (as the OED records): I have a set of desires (wishes, inclinations, commitments, etc.) and I can either act in accordance with them or against them, thus acting freely or not.[1] This has nothing to do with being free from all prevailing conditions: indeed I am not free from my desires (which may causally determine me to act as I do), but that doesn’t mean that I can’t act in accordance with them! I am free to follow my desires, because not prevented from doing so, even though I am not free from them. I may sometimes not be free to follow my desires, if I am imprisoned or shackled or subject to physiological upsurges that prevent me from acting as I wish; but much of the time I am free to do pretty much as I please (but see below). Quite often I am free to do exactly as I please, with no impediment at all to my freedom to do as I please. This is in no way compromised by my lack of freedom from antecedent conditions. Freedom-to is just a different concept from freedom-from; the locutions have quite different meanings and conditions of application. The compatibilist is thus right to insist that freedom-to is consistent with determinism, while the incompatibilist is right to maintain that freedom-from is inconsistent with the facts of historical determination. But the two theorists are not disagreeing with each other, once we distinguish between the two sorts of locution with their different meanings. The reason we feel the pull of each position is that both positions are perfectly correct so far as they go; we only get confused because we conflate the two concepts. And the reason we do that is that we yank the word “free” from its normal linguistic context and ask questions like “Does free will exist?” or “Is free will compatible with determinism?” Strictly speaking, these questions are ill formed, because they try to sever the concept of freedom from its surrounding grammatical context, which alone gives the word sense. We violate Frege’s context principle, or we fail to heed Wittgenstein’s warning about the perils of taking language on holiday. We are like someone who perplexes herself about freedom by trying to integrate the meaning of the locution “free with” (“John is rather free with his money”) with “free from” and “free to”. Is it that a free agent is one who is free with his actions? Can we be free with our past? Are our desires free with us? None of these sentences makes sense and can only generate pseudo-problems. Likewise, we should not try to shuttle between “free from” and “free to”, as if asking whether we are free to change our past or free from the future. In fact “free from” is a backwards-looking locution while “free to” is a forwards-looking locution: one connotes independence from the past; the other connotes dependence on desire in relation to the future. Am I free to act as my desires prompt me to? That is the question of freedom-to. Am I free from everything that has led up to this moment? That is the question of absolute freedom-from. And the answers are respectively: yes, I am free to act on my desires as opposed to being made to go against them; but no, I am not free from the conditions leading up to and surrounding my action, including my desires. I have freedom-to but I don’t have freedom-from. That is all that needs to be said, or can be said; there is no further question expressible as “Am I free?” or “Do I have free will?” It is not that I both have free will and don’t have it, or that I have to reject the plausible things said by the compatibilist and the incompatibilist; rather, I just have to return the word “free” to its natural environment tightly coupled with the prepositions “from” and “to”. Then (and only then) I will understand the import of our talk of freedom. The correct assessment of the philosophical upshot of this examination is thus twofold: (a) we are not free from our past, since our actions are determined by it; but (b) this does not rule out a robust sense in which we are free to act on our desires (the only kind of freedom-to there is). As a matter of fact, if we were free from our past, that would not provide an acceptable notion of freedom, since it would amount merely to randomness; and if we had the ability to act otherwise than our desires indicate (including our moral and prudential desires), that would not be a form of freedom-to. No occurrence in nature is free from the past, including human action; and nothing but acting on desire can add up to freedom-to. Nor is there any notion of freedom that is purer or better than freely acting on one’s desires—as if we are only really free when discarding our desires and acting in a vacuum.[2] For instance, a person who acts on his desire to save the world (perhaps putting aside his other selfish desires) is the paradigm of a free agent—and it is no impediment to this that his desire follows strictly from his genes and his upbringing. He couldn’t have acted otherwise, but so what—he was free to act on his most cherished desire. He was free to act on his altruistic desire despite attempts by others to thwart him, though his action wasn’t miraculously free from his mind and nervous system. The former freedom is not undermined by lack of the latter freedom.[3]

The difference between the two concepts is illustrated by a difference in their logic. An action is either free from a factor X or not; in particular, either actions are determined by antecedent conditions or they are not. It is an all-or-nothing matter. But forward-looking freedom-to is not so simple: a good case can be made that we are partially free in this way but not completely free. Am I really free to do exactly what I please, even in the most favorable conditions? Don’t I have all sorts of unrealistic desires that I can never act on? I would dearly love to fly like a bird, but I am not free to do so—the laws of nature prevent me from so acting. Don’t I also have a lot of conflicting desires that keep me from fulfilling all of them? Realistically, we can’t always do exactly what we please—we are not completely free. We are pretty free most of the time (if we are lucky), or more free than our neighbor, but we are not totally free. Freedom-to is not an all-or-nothing matter, unlike freedom-from. It operates in different conceptual terrain. It doesn’t breathe the same air. The logical behavior of “free to” is not the same as that of “free from”. This is why the compatibilist and the incompatibilist often seem like they are talking past each other: for they are talking about different things. The word “free” crops up in their discourse about these things but not because they have an identical subject matter—any more than its occurrence in “free with” discourse (compare also “tax-free”, “free as a bird”, “free society”, “free radicals”, “degrees of freedom”, “stimulus-free”, etc.) We mustn’t mix language games; we mustn’t tear “free” free of its linguistic auxiliaries. We mustn’t confuse one concept with another. Then we can accept that we have plenty of “free to” freedom but zero “free from” freedom—though remaining wary of that dangling use of “freedom”. Both are legitimate uses of the word “free”, but the constructions in which they occur have quite different import.

The intuitive idea of determinism is that the future is bound by the past, not able to escape its clutches, its shackles. This conflicts with the idea that we are free from the past and hence have alternative courses of action open to us. Thus we don’t have this kind of freedom. The intuitive idea of voluntary action is that we are often able to act without constraint or interference from sources external to our own desires, wishes, inclinations, preferences, values, and so on. This in no way precludes our actions from being genuinely free: to do what you want because of what you want is the very essence of freedom. Not indeed freedom-from, since we are not in so acting free from our desires (values etc.), but freedom to follow these desires without external impediment. Both lines of thought are perfectly sound: but, contrary to traditional thinking, they are not in conflict with each other. The plain fact is that we are free (to) but we are not free (from). I would recommend never using the word “free” in philosophical discourse without its attached preposition. That will make us free from confusion and free to stop worrying about the problem of freedom.[4]

 

[1] The fundamental idea of free action is surely freedom from other people: it is doing what you want irrespective of the wishes and actions of others. Internal factors can operate like other people, but the basic idea is that of interpersonal constraint or restraint. This has nothing to do with determinism; it is purely a matter of being free to act on one’s desires independently of others. So the notion of freedom is a social notion at root: if there are no other people, the question of freedom cannot arise. It is other people who put one’s freedom in jeopardy, not the past or one’s internal physiology or the laws of nature or one’s own desires. Have philosophers and others succumbed to a kind of anthropomorphism about such factors, modeling them on interfering human agents? That would explain a lot.

[2] It is sometimes said, quite correctly, that we are not compelled to act on our desires—we can resist their urgings and refuse to act on them. However, this is so only because we have other mental states that countermand these desires, typically commitments to values that conflict with the desire in question. These may take the form of second-order desires to the effect that the first-order desire ought to be resisted. The fact that a certain desire may incline us to action without compelling us should not be converted into an argument against psychological determinism, let alone physical determinism; for the reason for resistance will itself be another desire, possibly a value judgment, or some other psychological factor.

[3] It is instructive to consider free animal action. We recognize the difference between a caged or bound animal and a free-ranging one: the difference is the difference between the animal acting as it desires and being prevented from so acting. To be sure, its actions are determined by its genes, upbringing, and impinging stimuli, but that has nothing to do with the distinction between being caged and being free-ranging. Animals can be free to act on their desires or not so free, but they are certainly not free from the antecedent state of the world—they must act as they do. Still, there is all the difference in the world between being free to act on their desires and being coerced in various ways. The case is precisely analogous in the case of humans.

[4] I have been fretting about freedom for over fifty years and have wavered between different positions, most recently favoring the compatibilist position. This is my attempt to lay the subject to rest, at least so far as I am concerned.

 

 

The Selfless Machine

 

In a striking passage from The Selfish Gene Richard Dawkins writes as follows: “Other replicators perhaps discovered how to protect themselves, either chemically, or by building a physical wall of protein around themselves. This may have been how the first living cells appeared. Replicators began not merely to exist, but to construct for themselves containers, vehicles for their continued existence. The replicators that survived were the ones that built survival machines for themselves to live in. The first survival machines probably consisted of nothing more than a protective coat. But making a living got steadily harder as new rivals arose with better and more effective survival machines. Survival machines got bigger and more elaborate, and the process was cumulative and progressive.”[1] He goes on to say of these encased replicators: “They are in you and me; they created us, body and mind; and their preservation is the ultimate rationale for our existence. They have come a long way, these replicators. Now they go by the name of genes, and we are their survival machines”. This is the core of the “selfish gene” perspective: animal bodies (and minds) are the selfless conveyors of selfish genes into future generations. Organisms are artifacts constructed by genes to get themselves propagated. Dawkins is obviously thinking of the survival machines consciously constructed by humans (but obviously not consciously constructed by genes): houses, clothes, weapons, body armor, heating systems, etc. We construct artifacts that aid our survival, and genes do much the same. Their survival machines are squishy, organic, and sometimes conscious, while ours are dry, mechanical, and unconscious (so far at least): but the logic is the same—survival-enhancing devices. He could equally have spoken of survival kits or suits or sheaths or vehicles or pods or crafts or envelopes—the idea is just that of a complex entity that contains the replicators and enables them to reproduce safely. Thus a bodily trait is functionally just like a spider’s web or a beaver’s dam: it is a device for increasing the probability of gene survival in a competitive world. The spider’s genes are machines for making other machines (spider bodies), and these machines make still other machines (spider webs). The genes are the architects and beneficiaries of organisms, which in turn are the architects and beneficiaries of bits of external machinery (nests, burrows, mounds, tools, etc.) Genes are indirectly the producers of these bits of machinery, because they build organisms that build these bits. The external machinery is not selfish, as organisms are not selfish (in the technical sense of “selfish”); these things work to aid the selfish genes. They are selfless machines working for selfish genes.

If we add the extended phenotype to this picture, we can say that the genes build survival machines that incorporate both bodies and the external products of bodies (as well as minds in some cases)—webs as well as spiders. It might seem that this is a rather limited class of cases, since many animals don’t create anything tool-like. But closer examination reveals that actually extended phenotype is the rule not the exception, since animals modify their environment in myriad ways the better to aid survival. Many animals dig holes and these act as useful survival devices (so do plants with their roots). Animals tend to prepare their food before ingestion by chewing, chopping up, ripping, or infusing saliva. Birds use twigs as benches as well as build nests. Snakes use friction between themselves and the ground to propel themselves forward. Cats use spatial proximity to catch prey.[2] Fish use water to push their fins against. Organisms take advantage of the environment external to them in order to aid their survival: sometimes they actively construct useful artifacts and sometimes they find them already in existence. The survival machine includes all the facts that promote gene survival, extending out into the environment. It is arbitrary to locate a cut-off point at the animal’s epidermis. Put differently, the genes must take into account the environment external to the organism as well as to the organism itself when building a good survival machine. Natural selection favors good webs, but it also operates with hospitable found objects: it selects organism-environment combinations. A house includes not just the walls but the spaces in between, as well as proximity to water, etc. The phenotype is always extended in some way. In any case, external machinery should be reckoned to the machines the genes have created, as suggested by the extended phenotype. The phenotype survival machine built by the genotype is an extended phenotype survival machine.

So far this is pure Dawkins, stated in other language: it’s all selfish genes, selfless survival machines, and extended phenotypes. Now I want to venture further afield and push this perspective a little harder. The first question is whether the gene is the end of the line: might genes be survival machines for something else? Webs are survival machines for spiders, and spiders are survival machines for genes, but are genes survival machines for another type of entity? Do they function so as to keep something else in existence? Does their survival promote the survival of an even more basic biological entity? What about the chemicals that make them up? If a chemically complex gene survives, so does its chemical components—they become more common than potential rivals. So maybe the components of genes are entities that benefit from the survival of whole genes—selfish gene parts. After all, if a human machine proliferates, so do its parts: the more BMW cars there are in the world the more BMW car components there are in the world. The preservation of whole cars leads to the preservation of parts of cars (say, a particular type of hood). Thus we arrive at the idea of the selfish molecule. Second, isn’t it at least logically possible that we don’t know everything about genes, and lurking within them is some new kind of entity that shapes them? Maybe there exist more basic biological units that construct DNA sequences, so that the survival of DNA sequences ensures the survival of the more basic units. There is no reason that I know of to believe this is actually true, but its mere logical possibility shows that the gene is not necessarily the end of the line. Physicists keep discovering new and more basic particles: is it inconceivable that geneticists might discover even more basic genetic units? But third—and this is the point on which I wish to place particular emphasis—can’t we say that genes are the product of the process of natural selection, so that that process is the most biologically basic reality? For genes enable that processto survive: no genes, no organisms, and so no process of natural selection. Natural selection builds genes and without them it—that process–cannot exist: its very survival depends on the survival of genes. If natural selection started producing dud genes, ones that fail to build bodies that get the genes propagated, then the whole process would grind to a sickening halt—with nothing biological left. There is no process of natural selection on planets that contain no genes (or some equivalent): the process needs genes (replicators) if it is to gain a foothold. Looked at this way the genes function as devices for keeping natural selection around. So natural selection had better build good genes or it will go extinct. If we picture natural selection as a tinker whose livelihood depends on good gene tinkering, then an incompetent tinker is not long for this world—he goes to Valhalla if he produces bad tinkering work. Similarly, the process of natural selection depends for its survival on good gene construction (the kind that leads to robust organisms): the genes act as survival machines for that process. Genes are artifacts of that process, as organisms are artifacts of genes, and as tools are artifacts of animals. So the chain of production leads back to natural selection: it is the driving force of evolution, and genes act as its survival-enhancing devices. Natural selection is thus the ultimate “selfish” biological reality. To put it anthropomorphically, the genes are the indispensable servants or slaves of the process of natural selection: if they fail, it fails; if they die, it dies. The more genes there are the more natural selection there is, as the more organisms there are the more genes there are. It is in the “interests” of the process of natural selection that genes should survive, as it is in the “interests” of the genes that organisms should survive. We have existence-dependence at multiple levels: the spiders need the webs to survive, the genes need the spiders (and the webs) to survive, and natural selection needs the genes to survive (as well as the spiders and the webs). There is thus a whole hierarchy of survival machines terminating in the process of natural selection. Natural selection is the biological reality—the ultimate basis of the whole shebang. To paraphrase Dawkins, the process of natural selection, which may have looked paltry in the beginning, has gone far in its multi-million year history on earth, producing some remarkable survival machines. The DNA molecule itself is an impressive vehicle for the continuance of natural selection—it has enabled natural selection to colonize every corner of the planet. Natural selection now flourishes everywhere, while millions of years ago it was relatively confined and primitive. The gene was one of its greatest inventions, providing an excellent vehicle for it to travel far and wide. It is what has multiplied and proliferated beyond all imagining, aided by the trusty gene: without the gene (specifically DNA) it might have gone nowhere. If we think of natural selection as a species of process, comparable to species of chemical processes, then we can say that it is one of the most widespread species of process in the world. The genes are proud to have acted as its survival method. By building better bodies they improved its prospects dramatically by improving their own prospects. They gave natural selection a chance to operate widely and ingeniously; without them it would be eking out a living at a very low level of replication. The genes have maximized the spread of natural selection—as organisms have maximized the spread of genes. The logic is the same in both cases, though the entities are very different. Genes are complexes of molecules while natural selection is an abstract process. Complex molecules have enabled an abstract process to become an entrenched feature of planet earth. It is almost as if natural selection consciously designed them with that purpose in mind—as it can seem like the genes were conscious designers too. But no, in both cases we just have a mindless mechanical system following its own inevitable logic. It all follows from the very nature of natural selection as an abstract creative process.

Now that we are expanding the conceptual scheme of foundational biology let us revisit the extended phenotype. That was a good insight into the arbitrary nature of what might be called narrow phenotype—the kind limited to the animal’s body. The functional unit is really the body plus its adaptive effects (spider plus web); we could even reckon the web to be part of the extended body of the spider—it is that body spread out a bit further. But what about the genotype—must it be understood narrowly? Is the genotype confined by the boundaries of the DNA? It is telling that we regularly speak of genes by referring to their effects: genes for the kidneys, genes for the heart, and genes for the brain. We identify the gene by reference to the phenotypic trait it produces not by reference to its molecular composition. So why not work with the idea of the extended genotype? Why not take the bodily trait the gene produces to be part of the gene, as the web is part of the spider (i.e. its phenotype)? Not indeed part of the chemical composition of the DNA molecule, but part of a functional unit comprising molecule and bodily trait. Dawkins speaks of the “long arm of the gene”: yes, and it extends to the body it constructs in embryogenesis. The organism has a long arm, reaching to its adaptive artifacts; well, the gene reaches out that way too, to include its adaptive phenotype. Let’s call the external artifacts of an organism its “exotype”: then we can say that phenotype includes exotype and genotype includes phenotype—which includes exotype. By transitivity, genotype includes exotype—the web is part of the gene! That is, we reconfigure the boundaries of the gene so as to incorporate the external effects of the gene: the wide gene, as we might call it. There is the narrow body (spider alone) and the wide body (spider plus web); likewise there is the narrow gene (DNA) and the wide gene (DNA plus phenotypic trait). The canonical form of a wide gene description is thus “gene for X”, where X is some adaptive trait of the body or its products. It is the functional unit that is selected for by natural selection—a molecule plus the trait produced by that molecule (with accompanying apparatus). We thus insert the survival machine into the gene, widely construed—just as we insert the survival machinery of the exotype into the phenotype, widely construed. The web becomes part of the spider; the spider becomes part of the genes, or its traits do. We still have the narrow gene and the narrow spider in our ontology, but for theoretical purposes we also recognize another level of description that blurs such boundaries. If we imagine another biological world in which the same molecules are coupled with different traits, by virtue of different laws obtaining in that world, then we can see the point of carving things up this way: for in this world the same narrow genes will have different phenotypic expression, which obviously affects survival value. The same narrow gene will be adaptive in our world but not in this different world, given that it produces different disadvantageous traits. But if we individuate the gene widely we can say that adaptive value is preserved, since these are different genes in that world: the body comes along with the (wide) gene. What types of genes exist in a given biological world? Are they chemical types or trait-dependent types? If the latter, we only have the same genes when the traits are included in the total package—the extended genotype. Given that we are already comfortable with the extended phenotype, I see no reason why we shouldn’t accept the extended genotype (while retaining the narrow genotype—the gene without its long arm).

The initial replicators built their survival machines and even added external devices to enhance their survival capabilities. But later they recognized that they had merged with their machines and become extended beings. It is the same with spiders and webs: the initial spider built its web machinery and then stood proudly back admiring its handiwork. But later it recognized that it and the web were parts of a larger whole—it had merged with its web. The web survival machine became part of what the spider is—not significantly different from its legs and jaws. Similarly, the replicators were not just sitting snugly inside a survival machine but had joined forces with it: they were made up of bodily traits as well as tightly localized chemicals. To avoid confusion we might introduce a new term, since “gene” has become so strongly associated with spatially confined chemicals: we could call the entire complex of chemical material and bodily trait the “trene”, a combination of “trait” and “gene”. So trenes get selected by natural selection, just as combinations of spiders and webs get so selected. Trenes constitute the extended genotype. Organisms are collections of trenes, i.e. DNA molecules plus their associated bodily traits. A given trene would include, for example, a specific molecule in the spider’s DNA, a set of anatomical characteristics, and a distinctive type of web: this is what gets selected by natural selection (“selective holism”) not each of its components alone. Reverting to the point about natural selection as a beneficiary of the genes, we can say that natural selection builds survival machines with three main components: DNA, bodily traits, and external artifacts. Thus the process of natural selection survives because it builds machines that incorporate these three elements—that is, collections of trenes. Natural selection keeps going, keeps proliferating, because this is a good design for a survival machine—better than just bare replicators and better than organisms that can’t make anything. We ourselves, then, with our DNA, bodies, minds, and artifacts are vehicles for the continued existence of the process of natural selection.[3] We have become accustomed to thinking of ourselves as survival machines for our genes, but actually we and our genes are survival machines for an underlying biological process. We exist in all our glory because natural selection built us so that it could survive and flourish. We are the machines needed to allow natural selection to keep a foothold. We thus join the vast array of organisms on earth, from plants to people, bacteria to bats, whose job in life is to keep natural selection going. We are the devices invented by this abstract process so that it can survive. Natural selection created the gene (a certain type of replicator) and the gene created us (a certain type of reproducer): everything in the biological world is a survival machine for mindless natural selection to remain in existence and expand its domain. This is the ultimate rationale (to use Dawkins’ words) for the whole biological world. If you thought gene survival was desiccated enough, then natural selection survival surely brings desiccation to a new level. Even the mighty gene is made to feel secondary in the great scheme of things.

 

[1] The Selfish Gene (1989), pp.19-20. I am going to assume the perspective of this book in what follows.

[2] We could also say that cats use the device of injury in order to secure their prey: their teeth construct an injury that spells the demise of the prey animal. In the case of many big cats they use the windpipe of the prey as a device to subdue it, which is not fundamentally different from throwing a lasso around its neck. The environment takes on a function for the cat in its pursuit of food.

[3] This is not to say that we are only that—which is why I didn’t say we are just such vehicles. This is just one of the things we are, though quite an important thing. We are also moral beings, creative forces, free agents, romantic partners, scientists, etc. But from the point of view of biology we are the survival machines (conscious ones!) of the impersonal process of natural selection. This is ultimately why we exist.