Archive for July, 2009
In my previous post, I described the importance of a significance criteria for function ascriptions. My primary example was dueling function ascriptions for the Milky Way Galaxy. I argued that a natural function ascription for solid matter in the Milky Way Galaxy was a more significant explanation for the existence of solid matter, and the Galaxy, than an ascribing a biocentric (artifactual) function to the Galaxy. This argument rests on a presupposition that I did not explore however– that it makes sense to explain the functions of natural, but inorganic, entities at all.
Peter McLaughlin, whose What Functions Explain has spurred this series of posts, does not think that this presupposition of mine makes sense. Or at least that is the message I take from his late discussion of inorganic replicators in his book. McLaughlin writes that, “it makes no sense to speak of the function of some molecular substructure of a crystal, but we are not committed to the belief that the crystal somehow ‘benefits’ from being replicated” (2001, p. 181). Partially submerged in both of my previous posts on functional explanation is the idea that the split between natural and artifactual functions is a bit of a misnomer. In fact, the split seems to be much more between artifactual and organic functions, and so discussing functional explanation in terms of natural/artificial instead of natural/organic is a mistake. Merely being natural does not mean that something has a function, according to McLaughlin, but all organic entities do seem to have a function. McLaughlin seems to recognise this problem in his last few chapters, as once he has restricted the possession of functions to self-reproducing systems that have goods he begins to talk more in terms of organic entities that have functions and less in terms of natural entities that have them. To a degree, this problem is about semantics– if you don’t mean to include everything that falls under the category “natural”, use a different category, like “organic”!– but in a more important sense, the semantic problem reveals a real issue. Why don’t inorganic but natural entities have functions?
I think the split has been drawn between natural and artificial not just because it has a nice, dualistic ring to it, but because the terminology actually captures an important topic of metaphysics. There is a real difference between what it means for an artifact to have a function and what it means for a natural entity to have one. And, to add the kicker, that includes inorganic, but natural entities, such as the Milky Way. In my previous two posts, I have borrowed from McLaughlin function definitions that he ascribes to other authors, especially C.G. Hempel. But in the last (and very impressive) chapter, McLaughlin lets loose with a definition of his own. Answering the challenge posed in the book’s title, McLaughlin writes that functions explain “the existence and properties of those parts of a self-reproducing system that contribute to the self-reproduction of that system. What functions explain is systems whose identity conditions consist in the constant replacement, repair, and reproduction of their component parts” (2001, p. 209). This definition is meant to encapsulate organic functions only— artifactual functions are relative to the valuations of valuing agents, and (to repeat) it just makes no sense to think of the parts of inorganic systems as having functions.
I maintain that they do, and with alterations even McLaughlin’s definition of functional explanation can be made to accommodate them. Consider again my example of a system of extremely large scale– the Milky Way Galaxy. Then, expand the scale even further to that of the universe. According to most currently accepted cosmological theory, the universe is expanding, and has been since the Big Bang. Now, take the first half of McLaughlin’s definition of functional explanation– that functions explain “the existence and properties of those parts of a self-reproducing system that contribute to the self-reproduction of that system”– and integrate that thought with the idea that the universe is expanding. The universe is expanding, and (barring a misunderstanding, which, given the subject, seems almost inevitable), this means that there are parts of the universe that contribute to that expansion. These parts are parts of a self-reproducing system (the universe), that by their very existence, contribute to the self-reproduction and expansion of that system. Note that this would not be the case if the universe was in a steady state or in the process of collapsing into a Big Crunch sometime in the future.
There are clearly some enormous, massive difficulties in the details of this example. We don’t know what the stuff is that is causing the universe to expand. Some (a very small amount) of that universe seems to be matter of the kind with which we are used to dealing– solids, liquids, and gases– but most of it is apparently other stuff. Because I am not an expert Cosmologist, I cannot hope to explain these details. The expert cosmologists themselves seem to be at a loss, as well, but at least they have some ideas of how to work on these problems. What is important for my use of this popularly scientific understanding of cosmology is this: the Universe, the largest natural, physical, and (largely) non-living thing that can be conceived, seems to be a system that has parts that have functions. Many of those parts, being the structure of the universe, are just like it– they are natural and non-living. Yet they seem to be function bearers, at least with the first half of McLaughlin’s definition in mind.
At this point, you might be wondering if the universe, as a system, satisfies the second part of McLaughlin’s definition. Is the universe the sort of system “whose identity conditions consist in the constant replacement, repair, and reproduction of [its] component parts”? The rub for non-living systems in the definition seems to be the criteria of “replacement” and “repair”, included by McLaughlin in order to 1) distinguish natural functions from artifactual functions and 2) generalise the causal feedback loop required by functional explanations from requiring natural selection (which itself requires inter-generational reproduction) to intra-generational processes like the gradual maintenance of organs in living creatures (McLaughlin, 2001, pp. 179-190) This itself is an adaptation of a Maynard-Smith’s conditions for life, namely replication, regeneration, metabolism, and growth (Maynard-Smith, 1986). Regeneration, repair, or maintenance– no matter how you formulate the condition, it does not seem as if it is the kind of thing that can be assigned, broadly, to most parts of the universe. Which is why I suggest that we jettison the repair condition.
My motive for this is that parts of natural, non-living systems do seem to have functions. For the most part, except for this one criteria, parts of such systems meet the requirements that must be met by function bearers. The reason that I think natural, non-living entities should be included as possible subjects of functional explanation is that such explanation can be useful in providing more information about the universe. We are paying a heavy explanatory cost by limiting (severely limiting!) functional explanation of natural entities to living entities. Most of the universe is non-living, and we need all the information about it that we can get.
We do have to go about this sort of explanation in the right way. To lower the metaphysical costs of functional explanation, as well as to make the explanation more complete, we should be careful to limit (if not eliminate) intentional teleology, as well as anthro- or bio-centric perspectives, from functional explanations. Appealing to causal feedback loops as a primary ingredient of functions is currently the best way to achieve this goal. It is also important to keep the distinction between natural and artifactual functions, as one has an internal teleology and the other an external teleology, related to that of an external valuing agent. The primary objection to jettisoning the repair condition from a definition of natural functions is that by doing so one not only brings in natural but non-living entities but also artifacts. How do I propose to keep the distinction between natural and artifactual functions while rejecting the repair criteria? By building in a significance criteria. To get the most information out of an explanation, we should employ the variety of functional explanation that has the most significance for the system being explained. For a bicycle, an artifactual explanation is the most significant, and for the universe, a natural explanation is the most significant. It is not necessary to build into our definition of function that there is a distinction between artifacts, living entities, and non-living entities. We lose much by insisting that only one kind of functional explanation is possible for artifacts, another is possible for living entities, and no functional explanation is possible at all for non-living, natural entities. If we are truly motivated to accept a definition of functional explanation with the greatest explanatory range and power, and the lowest metaphysical cost, then we should not require that the subjects of such explanation be living if they are natural.
Peter McLaughlin provides separate formulations of what qualifies an artifact as having a function and what qualifies something naturally occurring as having a function. Most arguments over functional explanation seem to center around the functions of naturally occurring entities, although the apparent necessity of separate definitions for natural and artificial entities possessing functions is not without controversy. The problem with the split between natural and artifactual functions is usually located at the split, but I think there is also a major difficulty when the two come together.
McLaughlin’s basic summation of natural functions is this:
(1) X does/enables Y (in or for some S)*
(2) Y is good for some S; and
(3) By being good for some S, Y contributes to the (re)production of X (there is a feedback mechanism involving Y’s benefiting S that (re-)produces X). (2001, p. 140)
And for artifactual functions, he provides the following schematization:
(1’) X does Y—or at leas some relevant agent believes it does.
(2’) Y is good for some S—or at least some relevant agent believes it is. (ibid)
Now, to muddy the waters a little bit. It would seem that natural entities and artifacts might be divided along lines of living and non-living things. This is not the natural entities that are living ones, but is to say that “artifact” seems to intuitively include only non-living things. After all, living things cannot be made into artifacts, right? More than that, natural things cannot be artifacts, right?
This being philosophy, the line is not nearly so clear. What about domesticated animals? On the other end, in the non-living but nonetheless natural range, what about the Milky Way Galaxy? There seems to be a large enough class of things that blur natural/artifactual distinctions that there might be a confusing proliferation of function ascriptions, not to mention difficulties in classification.
Let’s look at one of my two examples that blurs the artifactual/natural distinction, the one that is more extreme: the Milky Way Galaxy. How, you might be wondering, might the Milky Way be an artifact? Under the guidelines for artifactual function ascriptions, the Milky Way (and the rest of the Universe, for that matter) might be said to (1) provide a substrate for life and (2) that providing a home for life is good for humans. In fact, this is much of the fine-tuning argument in a nutshell. One immediate problem with this ascription of function is the insignificance of providing a home for life (at least to the best of our knowledge) to the existence of the universe. It is an explanation incredibly small in scope– to say that a function of a galaxy is to provide a home for life is similar to declaring that a function of mine was to touch a particular token at an arcade in 2004. It is to pick out a physically and temporally small occurrence in the history of an entity and hold it to be a function.
Another problem is that one can also cast a functional explanation for the Milky Way as a natural entity. (1) The Milky Way Galaxy fuses particles together in such a way that solid matter exists. (2) Solid matter is essential to maintain the shape and makeup of the Milky Way. (3) The more solid matter there is in the galaxy, the more of the galaxy there is, and without solid matter there would be no Milky Way. This certainly seems to be a more significant function of the Galaxy than providing a home for life, not to mention that solid matter seems to be more important to the actual continued existence of the Milky Way as such than the existence of life in some corner of the Galaxy. Except in a linguistic sense (without language, the Milky Way Galaxy would lose its name), there is not much that the continued existence of life does that maintains the existence of the Galaxy.
Before function ascriptions can be made for natural entities and artifacts, perhaps there first needs to be an additional guideline on how to decide which function ascription to employ. This could take the form of a significance criterion, similar to that described in Philip Kitcher’s The Advancement of Science (1993). There, Kitcher asserts that a goal of science is not simply truth, but significant truth. There are all sorts of trivial truths that are easy to arrive at, but science is after those that are significant. Similarly, to ask for an explanation of an entity’s function one is looking for the most significant of the range of explanations. While life is important to some features of the Milky Way, solid matter is much more so. Focusing on a natural function of the Milky Way explains more– it helps explain more of how the Milky Way exists at all– while an artifactual function ascription explains the Milky Way from a very unrepresentative biocentric perspective. This is not to say that the artifactual function of the Galaxy provides no information, but the natural functions of the galaxy are much more significant and provide explanations with greater depth.
I am pressing ahead with Peter McLaughlin’s What Functions Explain. Although I have more reading to do– and McLaughlin is judicious in his pacing, frustratingly raising interesting questions and promising to engage them in later chapters so as to continue on a single subject– his discussion of the etiological approach to functional explanation raised some interesting thoughts having to do with a contrast between certain practices in science and religion.
Let me go over that territory again before linking functional explanation to science and religion. Exactly what entails a functional explanation– or an explanation of some entities function– is difficult philosophical territory. In contemporary philosophy, there are two main competing schools on what a function should consist of, originating with Thomas Nagel and C.G. Hempel. The etiological approach to functional explanation relies on including a causal aspect in the explanation. Instead of a nomological statement or two– along the lines of:
The function of an item X in a system S is to do Y, which under some condition C (including internal state Ci and environmental context Ce) is required for, or is at least conducive to, goal G. (McLaughlin, 2001, 68)
The Hempelian variation on this theme is broken into three parts by McLaughlin, namely disposition, welfare, and feedback. Disposition is the tendency of something– an organ, artifact, organism, or other entity– to do something under some condition (as above). The welfare condition is that that something tends to promote a goal of that entity. Finally, the feedback loop is the etiological (causal) characteristic of the explanation. It relates to how the entity has come to perform its function and how the performance of that function is vital to the continued existence and performance of that entity. In other words, it isn’t enough for a functional explanation to feature a means and an ends: it must also feature how the performing entity came to be a function bearer, and continues to exist as such. It must be part of the function of a thing to put it in place to perform its function in the first place, instead of its coming to perform its function by accident.
Naturally, this account raises all sorts of problems. I will take McLaughlin at his word, though, and continue on in his text before I take the plunge into that discussion. For now, I’m going to ignore some of the problems with the etiological account of functional explanation and go forward with how such explanation is illuminating for a description of science and religion. At least some of the time– when there aren’t issues with the vagueness surrounding the term “function”– I think scientific explanation makes use of this kind of explanation. In particular, the issues scientific explanation would have with function seem more oriented towards the dispositional and valuational characteristics of functions and not the etiological aspect. The opposite seems true of religion.
One of the issues, of course, is that of teleology and its attendant baggage of intentionality. The more interesting issue is that of the causal feedback loop. The issues scientists have with teleology and its gestures towards an unknown agency are fairly well-known, and so I will leave them to other authors or separate posts. What caught my interest today are the different demands of science and religion on functional explanation. Science seems to demand an etiological component, and religion does not. This is because of the two different approaches science and religion have towards causation. Religion, (if the definition includes some content regarding the supernatural) is not merely permissive of interrupted causation in the natural world– it is demanding of it. Science, on the other hand, stresses a uniformity of causation in the universe.
The significance of this is that causal feedback loops may be an essential ingredient of an adequate scientific explanation, while they create a source of tension with some religious explanation. A causal feedback loop provides closure in an explanation– the heart is frequently used as a classic illustration of this. The function of the heart is to pump blood, which certainly contributes to the welfare of the organism that contains the heart. As long as the heart continues to pump blood, the organism will continue to live and cause the heart to pump blood. Short of some interruption in the system, the heart will continue to function. The heart was developed particularly to do this, in that sort of system, and will continue to do so unless interrupted or replaced.
If one insists on a more intentional world (at the extremes, this consists of belief in special creation), this explanation might seem incomplete or unsatisfactory. “Why,” such a person might ask, “did the heart come into existence at all?” This question is intended not to probe the intricacies of the causal reinforcement loop– the reply, “it developed to pump blood” is really only a restatement of the information contained above, and it is not the sort of answer the questioner is after, anyway. This is because such a question is intended to step beyond the causal loop, and beyond the uniform sort of causation contained within. Doing so does away with the need for the etiological condition, reducing functional explanation to the dispositional and the valuational while providing an overlaying agency to the world. Now, the explanation for the function of the heart is this: The function of the heart is to pump blood, which certainly contributes to the welfare of the organism containing the heart. The heart came into being with organisms that required them to, themselves, function. It is not difficult to imagine this line of thought terminating in something akin to Aquinas’ arguments, a sort of regress that ends in the divine. This is a familiar sight in philosophy– if unsatisfied with a loop, the other choice is a regress. What had not occured to me before is how scientific explanation tends to be comfortable with the one, and religious explanation the other.
Since my last posting of my proposed reading list, I’ve received a ton of suggestions. So many that, at the rate that I also discover new and interesting books, I probably won’t get to them all by the end of the summer. Currently high on the list are Yukio Mishma’s short story collection Death in Midsummer and John Locke’s Essay on Human Understanding. As early modern philosophy is a huge weakness of mine, I figured now was a good time to shore it up with some good, old-fashioned British empiricism. Also, I’ve been meaning to read his views on substance and qualities since I took “Philosophy and the Environment” at Edinburgh last year. Because that University had an understandably old-school approach to philosophy amongst its undergraduates, all of the students in the class that WEREN’T exchange students seemed to follow such Lockean discussion while I barely held on.
Of course, I still need to finish Orientalism and What Functions Explain. Reading some of Kitcher’s essays would also probably be time well spent. And I picked up an essay collection on Locke, hoping that it would offer more clarification than confusion. In the land of fiction, I found Pynchon’s Gravity’s Rainbow on the side of the road, so that might bump up after I finish Mishma’s stories. I also would like to read Snowcrash, The Brothers Karamazov, and all of your recommendations. We’ll see how I fit that into my schedule that involves my heavily resembling a french fry for about 8 hours a day. Awesome.
Rouse Up… is ostensibly about a famous author father, K, and his relationship with his son, nicknamed Eeyore. The book, like many of his, is semi-autobiographical, based on Ōe’s own life experience with his disabled son. I say “ostensibly” because it is not really about Eeyore at all. Told in the first person by K, the story is primarily about the narrator. Other characters come in and out of his life and thoughts like planets with elliptical orbits. Eeyore is the most prominent of these satellites, and the one that leads to the most reflection, but the story is not about him. The most that can be said of Eeyore is that his inscrutability, caused by his personality and amplified by his disability, leads K to meander in his own thoughts, trying to decipher ways to explain life, death, and imagination to Eeyore while at once attempting to discover if Eeyore already understands them. For almost all such concepts, it is revealed at climactic moments that Eeyore does understand those topics that K thinks most valuable, but has done so organically, largely independently of K’s own parental and literary artifice. The benificary of such work, of course, is K himself.
This is not to say that Rouse Up… consists solely of unself-conscious projection: Ōe has a deft touch as to the timing of K’s revelations, in terms of K’s criticism of himself, his parenting, and his place in society. Sometimes these moments occur, as I said, in tandem with an event or phrase of Eeyore’s, sometimes within frequent meditations of the work of William Blake and its applicability to K’s own life, and sometimes on one of many tangents and switchbacks into K’s life and travels. The last device, whose use sometimes makes it difficult to follow exactly how events in the main plot are unfolding, is what makes Ōe fit with Kawabata and Mishma. Rather than plowing through events involving K, Eeyore, and their family in a strict, chronological narrative, K is led to describe episodes from his childhood, education, and professional life. Early on in Rouse Up… this is done under the pretext of explaining K’s view towards life or death so as to find some way to explain them to Eeyore, but as the narrative goes on the episodes are introduced more and more as personal reflections performed for personal moments. One theme in the story is the frequent criticisms levelled at K that he hides behind Eeyore in order to never change, progress, or take a stand on modern issues. Ōe answers these charges, (which, given their prevalence in the story, have likely been leveled against him), by allowing his narrator to mature in front of our eyes.