Find the Pope in the Pizza

When I espy the vast canopy of the nighttime sky, the solemn majesty of the Grand Canyon or the brooding presence of the Himalayas, when I see a green shoot thrust its way from the late spring branch, when I see kestrels in flight or lion cubs at play, the first thing I think is, “I could really use a fuck right now.”

But let’s assume I’ve already had three orgasms today and that I’m tapped out in that department. It’s time to do some philosophy. I have dealt elsewhere with the logical problems of the concept of a designer or creator god, the supposed solution to the apparent design in the universe, using as a touchstone Richard Dawkins’ disproof in The God Delusion. I have also dealt with one objective index of design, namely the apparent improbability of certain properties of the physical universe. I considered statistical arguments for improbability and also the so-called anthropic principle. Now I would like to turn to the ultimate supposedly irreducible assertion on the part of those who claim to perceive design in the universe and examine what it is they actually see. (A simpler version comes from those who claim to see god somewhere, or more frequently talk to this character. The immediacy of their “observations” is supposed to have the same unassailable character as Moore’s conviction that at some time in the past he saw a white patch. Both the design perceivers and the god perceivers come down to saying that no one can argue with what they see or hear.) Does it make sense to say that the universe looks designed? Are alternative descriptions really excluded? Despite the fact that actual design gives every indication of being a red herring and suffers from serious logical problems, the appearance of design does tie in with truly interesting topics in the philosophy of perception, the nature of scientific verification and even artistic taste. And there are the practical and moral issues. Remember the likely reason you can’t have a threesome with that hot blonde and her girlfriend is that some goddist sees design in the universe.

The scenario is a familiar one. A simple woman of faith (perhaps named Maria, since most simple women of faith seem to be named Maria), after a long day of cleaning houses, drags her weary body to the local Pepe’s Pizza and orders a 24-inch thin crust with extra sausage. When she gets home and opens the box: Behold, etched in the sausage is the photographic (or as photographic as sausage can get) image of Pope Armando XVII, Maria’s favorite pontiff since the day he seemed to smile at her from the TV, his gold tooth glinting in the afternoon sun. Well, no dinner for the kids (all eighteen of them) that night. Maria runs (or rather waddles) to the neighbors with the glad tidings and those who don’t immediately kneel in vigil before the pizza box make a beeline for the local priest, Father Guido Sarducci. The rest is ecclesiastical history. Journalistic helicopters swooping overhead, the nation’s legislators opine that a simple pizza box provides the ultimate refutation of Communism. Rumors swirl that the pizza spoke to Maria predicting the end of the world. She is said to be a candidate for beatification.

Design theorists are academic Marias, gorping at watches in meadows, worshiping fantasy pizzas and expecting the rest of us to take them seriously. The pizza looks designed (Well, it was designed, but by Pepe and not by god), and since that clever clue looks designed, everything else must be designed, and so god… (Do I have to drag this out?)

Like the design theorist Maria argues from perception. Her argument has minimal deductive content. When you use an argument from perception to convince me of something, you rely almost exclusively on the fact that I will see what you see. For that reason it appears to be very difficult to defeat since how can I show you that what you think you see isn’t there or that you don’t see what you believe you see? The evidence is in front of your eyes. Well, let’s see what we can do.

The claim to “just see” design in the universe is the last redoubt of the defeated creationist. Even Philo subscribes to this argument-ender, though not perhaps without a dash of disingenuousness on the part of his creator for whom a snub down at the country club was the moral equivalent of Bruno’s auto da fe. Let’s temper Hume’s unconvincing volte face and preface the discussion of perception with a summary of his demolition derby of anti-design arguments in the locus classicus of atheology, the Dialogues Concerning Natural Religion.

I. Hume’s Arguments

1) The Weakness of Analogy. Philo (p. 46) uses analogy broadly to include all of what we would call inductive reasoning. In fact Hume is apt to regard any reasoning that is not either purely logical in his sense of “logical” or mathematical, as he understood “mathematical,” to be analogical. Analogies do not constitute proofs. Rather they are explanations of our inclination to believe. We conclude that the sun will rise tomorrow by analogy from previous mornings. But our conclusion is not the result of a proof. This is a much more important difference than it may appear to be at first glance. Hume modestly qualifies this point as a prime example of what he calls his “moderate” skepticism. In fact it represents a significant shift in the way we understand reasoning in the empirical or non-logical sciences. Reasoning by analogy, Hume observes, does not have the force of logical proof. But if you look closely at induction interpreted as reasoning by analogy, it is not even like a proof. It is a psychological event and a theory of analogical reasoning is a psychological theory. What Hume actually proposes is a motive as to why we should draw certain conclusions. But this motive contributes nothing as to judging whether our conclusions are right or wrong. Hume’s theory of analogy is itself a brilliant empirical theory, on a par, let us say, with natural selection. Absent any other reason assigning validity to a given induction, conclusions from the empirical sciences may not be amenable to proof in the rigorous sense at all. Only the deductions embedded in the chain of reasoning leading to the conclusion have a proof-like structure.

Nevertheless there are, according to Philo, stronger and weaker analogies. The basis for discrimination is the strength of the similarity between analogous events or phenomena. The strongest possible analogy is one where there is an “exact similarity” between events or phenomena. As the similarity diminishes the analogy grows weaker. It is a strong analogy, says Philo, to conclude from the discovery of circulation in frogs and fishes to circulation in other animals, but a much weaker one to conclude that circulation also exists in vegetables. Philo’s account leads straight back to the central issue. On what basis can we conclude that one similarity is stronger or weaker than another? What does it mean to say that we perceive a similarity? (Of course, Philo's description of analogy is an argument from perception or observation. But the point here is not to disqualify arguments from perception entirely; rather to distinguish valid from invalid applications.)

Design arguments constitute an overt form of analogical reasoning, for they are based on the assumption that the universe “looks like” the objects of human design and creation. The gist of Philo’s opening sally is that the universe looks nothing like an object of artisanal creation such as a house. The design argument fails because the basic premise is unconvincing.

There are weaknesses to this viewpoint. For one thing there is no mechanism within the theory of analogical reasoning itself that allows us to discriminate between strong and weak analogies. It merely says that, if we see like sequences of events with sufficient frequency, we are apt to conclude that such sequences will repeat in the future. Hume’s insight lies in connecting frequency in the past with future predictability. The weakness consists in deciding what constitutes a “like” sequence and what constitutes stronger or weaker likeness. Philo does not propose a method for deciding between stronger and weaker likeness. As an empirical theory his must rely on people’s past behavior and his own current intuitions so that he may venture to assert that one group of phenomena is more or les like another group. A related problem is that Hume simultaneously uses the theory of analogy as a pure empirical theory, a simple statement of certain facts about human behavior, on the one hand, and as a method for evaluating analogical arguments, on the other. As stated above, there is nothing in the empirical theory of analogy by itself to warrant judgments of the validity of specific inductions. A better version of moderate skepticism would most likely rely heavily on observation to determine the perceived strength of an analogy. The proper phrasing would be not which analogies are stronger or weaker, but which analogies have been believed to be stronger or weaker. Hume very well could have resurrected Aristotle’s famous, “…so men say.”

An important distinction that Hume fails to make is between induction or analogical reasoning as a spontaneous bit of behavior and induction or analogical reasoning as an expressly stated argument. Hume tends to treat all - including unconsidered - expectations as if each one were the result of a conscious review of past facts such as to draw an explicit conclusion. Much of what Hume appeals to are simple expectations and not arguments. But not all. An empirical theory like the extrapolation of the theory of the circulation of blood to all animals is an inductive argument. Also much of molecular phylogenetics consists of hypotheses drawn from analogical reasoning. The case is made by appealing to other points of similarity between known and unknown cases, similarities such as capacities for sensation and locomotion in the case of the circulation of blood and broader morphological and behavioral similarities as well as similarities in amino acid and gene sequences for phylogenetics. These other points of similarity, of course, are either spontaneously held to be true or themselves the result of further inductive arguments. This may provide a means of distinguishing between stronger and weaker inductions. Unreasoned inductive beliefs may be stronger than the conclusions of inductive arguments, not because they cannot be wrong (After all the flatness of the earth was a spontaneous inductive belief) but because they are facts; they themselves belong to the realm of observable phenomena for which the scientist tries to provide an explanation. Inductions that are the result of argument, however, by their very nature have no history of having been held to be true by anyone in the past. They must be established by an accumulation of inductive arguments that ideally ultimately appear to be credibly believed spontaneous inductions.

It is not a “refutation” of Hume’s theory of inductive reasoning to point out that it also is an empirical theory. Hume’s point is that, outside of logic and mathematics, there is no such thing as rigorous proof or refutation. His is an empirical theory that will serve so long as it provides a suitable explanatory model of the phenomenon of empirical proof.

On this basis it can be cogently argued that the design argument is an induction by argument and not a spontaneous inductive belief. One clue lies in our intuitions and current behavior. If we see a forest we draw a distinction about whether it is appropriate or not to ask who planted it. If all forests were strongly similar to cultivated groves, then it would not occur to us to ask such a question based simply on the appearance of the forest. Furthermore, this aspect of the design argument now becomes amenable to historical and philological research. For example, do primitive cosmogonies that assert that the universe is the product of an external will exhibit the quite distinct assumption that the physical universe shows signs of design similar to human artifacts? Does the replacement of mythical with naturalistic cosmologies accompany general changes in what are considered similar phenomena?

The appearance of design in the universe is itself a result of analogical reasoning and cannot serve as a “self-evident” basis for a conclusion that the universe was created. The inductions include the assumption that the universe is similar to a humanly created artifact such as a house, that all things that look like houses are houses or that all houses have been designed and not simply appropriated by protection against the elements (Some caves and forest groves after all are nigh perfect for the purpose). The weakness of the analogy in the design argument lies in the fact that it is not a widely held assumption (Remember the assumption in question is not that the universe was created by an external volition but that the universe “looks like,” “is similar to” a human artifact) such as the assumption of the approach of dawn. For that reason the assumption of similarity cannot serve as a secure basis for the analogical conclusion that the universe was indeed created by a human-like external volition.

An interesting historical fact is that Indian logicians of the so-called Cārvāka school argued against analogical or metaphorical reasoning disguised as syllogistic deduction as early as sixth/fifth century BCE (Cf. Mádhava Áchárya, the first chapter).

2) Order Does Not Imply Design. “Imply” here is used in the strong logical sense of material implication. An alternative formulation is, “Order does not presuppose design,” where "presuppose" is also interpreted as material implication. Philo (pp. 48-49) uses the Law of Non-Contradiction as his criterion. The proposition that the universe exhibits order does not contradict the proposition that the universe is not designed and vice versa. It follows that the design argument is not an a priori proof as Hume understands “a priori.” If valid at all it also is a type of analogical reasoning or induction. That is, only after several cases where order is demonstrably the result of design, where we have experienced the causal link, can we extrapolate to new cases. It is indeed possible that there is an “original principle” or “internal cause” in “matter,” to use Hume’s terminology, such that the experienced order of the physical universe is the result of that internal principle. Philo immediately backtracks and concedes that the majority of individual cases of perceived order that we have experienced are the result of human design and not the spontaneous activity of undesigned matter. Stones, steel and the rest will not arrange themselves into watches and houses. At this stage there may be some non-a priori experiential basis for extrapolating to the universe as a whole. Nevertheless, despite Philo’s concession, this famous passage sets the stage for the discovery of principles of ordering that are distinct from humanoid intention. In the biological realm at least, all that was needed was something like the theory of natural selection to cash in Hume’s promissory note. I might add that Hume was already sitting on top of another such principle for the inanimate universe in the form of Newtonian mechanics. Complex orderly (more accurately repetitive) phenomena are explained by appeal to a small number of properties that are related by mathematical ratios. Newton’s own theological ramblings might have obscured that fact, but Laplace would soon show that Newtonian mechanics could be adequately formulated without theological appendices.

3) A Humanoid God is Not Perfect. Philo quickly limits his concession by pointing out (p. 48) that the analogy between the artifact-human design relation and the universe-goddish design relation assimilates god to an anthropomorphic model. But humans are not perfect, so the suspicion is that god, thus understood is not perfect either - an unacceptable conclusion. One is tempted to demur at this point to the effect that it is not contradictory to assert that god is humanoid enough to have designed the universe and yet remain perfect, however vague such a concept of god may be. It is worth keeping in mind, however, that Philo is not proposing a disproof (certainly not an a priori disproof) of the existence of a designer god. He is rather evaluating the strength of the analogical argument. In order to maintain a strong analogy based on exact similarity, god must be conceived as sharing human imperfections. To the extent that god deviates from what we experience to be human, the analogy, and by extension the design argument, grows weaker.

One can argue (and I have done so elsewhere) that any concept of god or description of god that deviates from an extreme minimum invites problems of this sort. Though the problems need not necessarily arise specifically from anthropomorphism, most concepts of god and certainly those in the Maxju tradition are largely if not wholly anthropomorphic and drawn from predicates that are otherwise used to qualify human qualities, attitudes and behaviors. Demea (p. 57) raises Philo’s humanoid argument to plump for his view that human understanding is limited and cannot grasp god in its true ineffability. But if Demea is right, i.e. if we cannot criticize proofs or even talk about god based on the fact that those proofs involve the anthropomorphizing fallacy, then we cannot even propose the proofs or talk the talk. The jargon of ineffability puts an end to all god talk, not just assertions about god but also propositional attitudes toward god including worship, obedience or whatever else you want to throw in.

4) Part-Whole Analogies are Weak. The idea (pp. 49-50 and p. 79) is that the properties of a part of an entity do not necessarily transfer to the larger entity of which it is a part. Some engine parts are after all hotter than other parts. The springs of that favorite 18th century example, the watch, possess an elasticity that the hands and face do not. The same applies to having been created. Philo’s own not completely felicitous examples are that the growth of human hairs and leaves are disanalogous to the generation of humans or trees. These examples are not happy because there is a strong biological relationship between the generation of the parts of an organism and the generation of the organism itself. A better example might be a scene in which a primitive human takes materials from a cave and fashions some sort of tool such as a hammer. The hammer is designed and created and the materials are part of the cave but the analogy to the origin of the cave fails. When we see a cave we do not immediately ask who hollowed out the ground (Some goddist petitio principii here is not an adequate response. The weakness of the analogy is exhibited by the fact that we don’t immediately ask which human hollowed out the ground). Seeing the hammer most certainly does not inspire that question.

The part-whole argument is entangled with a quite distinct point to the effect that there are other principles affecting generation besides “Thought, design and intelligence such as we discover in men and other animals....”  There are other contributory “springs and principles” such as heat and cold or attraction and repulsion. This important and somewhat buried comment is a nod in the direction of Newtonian mechanics that I mentioned above. Design is distinguished from the other principles as an “active” cause. The important unstated question is: Must there always be an active cause for something to be the way it is? Is an active cause a necessary condition for any phenomenon? Hume’s answer would be that, as long as the denial of an active cause is not contradictory, no. (There are unexplored implications here for the tantalizing but perplexing view that the universe itself exhibits a kind of conatus or will.) In fact, dispassionately considered, there may not be a lot to be said about animal will or animal design as a creative principle. “This little agitation of the brain which we call thought…” is one empirically and possibly introspectively observed phenomenon among others “on this planet.” Indeed as an effective principle of change it is minute, weak and bounded.

5) Part-Part Analogies are Weak. This is clearly a variation of the part-whole argument. Philo’s illustration (p. 50) is less than definitive in view of modern developments since we now assume sufficient statistical likelihood of organic life and perhaps also intelligence in “remote” parts of the universe to encourage further research. That statistical likelihood, however, is based on empirical information, unavailable to Hume, of the conditions for life in other parts of the universe, information such as the widespread presence of water and organic molecules. Indeed the actual somewhat different point is still compelling despite the weakness of the illustration. There is almost no similarity between the creation of a component of the universe in its present state and the creation of the entire universe at some very distant point in time. The fact that houses very often have builders does not in the least intimate that whatever appeared at the origin also had a “builder.”

6) Inductive Inference to a Unique Occurrence is a Fallacy. A proper inductive inference, or what Philo calls an “argument from experience” (p. 51), is one based on the actual experience of numerous occurrences that were identical or very nearly identical (By “identical” I obviously mean sharing the same relevant qualities; not numerical identity) to each other. When a new situation comes up that is like those previous occurrences in all respects except one as yet undetermined, we expect that undetermined respect to be the same as well. When the famous billiard ball strikes another one, we expect the second one to move. The origin of the universe, however, is (almost by definition) a singular event. We have nothing to compare it to. We would have had to have experienced numerous origins of universes to have any basis for an inductive inference. The correct analogy is to occurrences of substantially the same type and not to dissimilar occurrences in a universe already in existence. (5) and (6) are very closely related arguments, but they are different.

7) Universal Consent is Not a Sufficient Argument. The universal consent argument differs from the basic design argument for the existence of god in that it does not appeal to a given individual’s perception of design in the universe. It is meant to answer the very strong objection on the part of an atheist that the universe does not appear designed to him. It is something like an argument from authority. Almost every culture has had some sort of religion and the leading theologians all see design in the universe. Cleanthes prepares the way (p. 61) for the refutation of the universal consent argument by noting that many if not most theologians in the Xtian tradition subscribe to the notion that god has no qualities or attributes. This is the doctrine of perfect simplicity. However, an entity with no attributes cannot even be talked about. Any reference to it has no meaning. At best the perfect simplicity doctrine is a form of skepticism and the theologians who hold it are atheists à leur insu. But, observes Philo (p. 62), the three interlocutors agree that “idolaters” are also atheists (The reason presumably is that the objects of idolatry, sticks and stones, are not god). Accordingly the majority is actually atheist. Appeal to universal consent proves the negative of what it intended.

8) Infinite Regress. This is the famous argument reprised by Dawkins to the effect that any cause of the universe itself requires a cause for the very same reasons we demand a cause for the universe. On the anthropomorphic model used by design theories, the cause of the universe must be something like a human mind. But, if the universe is ordered in a certain way – and it is precisely because the universe is perceived to be ordered in a certain way that a designer is hypostatized in the first place – then the designing mind must have an equivalent sort of organization. But, on the premise of the original argument, this organization requires explanation. We must search for a new “intelligent principle.”  But “How can we satisfy ourselves without going on in infinitum” (p. 63) Hume and Dawkins have a built in response to anyone who might say further designers are redundant. The point is not that we should go out and find a cause for the designer god; the point is that the initial hypostasis of a designer god lacks explanatory value in the first place. Hume says as much when he observes that if a cause of the cause of the universe is redundant, then so is the cause of the universe (Cf. also pp. 81 ff.). If god doesn’t need a separate creator, then neither does the “material” universe. I have argued elsewhere that infinite regress is not just unsatisfying and non-explanatory. It also involves serious logical errors that render elements of nearly any theory of a designer god literally meaningless. Not even false. Meaningless.

9) The Perfection of the Xtian God Means It Cannot be Qualified as a Creator. This argument (pp. 68-69) is a distinct variant on argument (3). It can be construed in two ways. The first is that a perfect entity cannot have created our universe, which is imperfect. The second - which would be valid even upon the supposition of a perfect universe - is that the creator of something may be inferior to what it created. That is to say, a created universe does not imply the perfection of its creator. But Xtian theologians and Deists have defined their god as a perfect entity. Defenses have been mounted against the first construction to the effect that a perfect entity may have its reasons for creating an imperfect universe or that our universe may really be perfect after all. Responses to these defenses have asserted that, if there was a good reason for creating an imperfect universe or if our universe has some sort of higher order perfection, a truly perfect entity could have a created a more perfect universe that included the perfection of our universe without the detour of the apparent imperfections. The prize in this dust up, which is clearly the argument from evil in different language, clearly goes to the side that sees incompatibility between a perfect creator and an imperfect creation. Any confusion in the matter stems from ambiguities in the concept of perfection. I argue elsewhere that the term “perfection” is largely meaningless, that like “being,” “good,” “all,” “freedom” “beautiful” and even “entity,’ it derives its fake meaning from playing around with words and arranging them in grammatically correct but semantically challenged patterns. They are constructed much like inconsistent (as I define “inconsistent”) terms such as “the Jenna who has beautiful nipples and who does not have beautiful nipples.” They are what I call fucked up terms, and understanding how they are fucked up and what our reaction to their fucked-upedness should be constitutes one of the main problems of fuckosophy. A further defense is often offered at this point to the effect that the universe and eo ipso its creator are really perfect. We humans are just too limited to see how they are perfect. This appeal is simply a reversion to the jargon of ineffability that I dealt with in (3) above. But it also suffers from a logical fallacy peculiar to itself. Since we, the observers of the purportedly perfect universe are part of that universe, we should perfect as well. Accordingly we should be able to perceive and understand its perfection.

Philo’s second form of the argument is meant to show that, even if the perception of design succeeded in convincing us of the existence of some sort of creator, that creator may very well be inferior to god as theologians understand god. The creator may be a “stupid mechanic” who copies someone else’s plans. It may have arrived at this state of the universe through a process of trial and error. It may be “several Deities,” a sort of team of creators. It may be married or a bumbling infant, or something really very much like Jupiter after all. Even if “something like design,” were to somehow be proved, no conclusion “beyond that position” can be drawn.

10) The Universe Does Not Even Look Designed. In Parts VI and VII Philo returns to the point he raised earlier (p. 46) and gives it more content. The scheme of the argument is, so to speak, polysemic. Hume’s explicit contention is that, to him at least, the universe does not look like a product of human artifice. It looks more like an organism: a vegetable or an “organized” human body. Arguments about how we see things or what we see lie on particularly thorny ground because, at one level, our perceptions are not themselves theories, but the bases of or the evidence for theories. You cannot even begin to defend the theory that the butler did it if there is disagreement about whether there was even a corpse or whether Sir James had actually been observed dead or living. Moreover, differing perceptions raise difficult issues of translatability. If someone appears to see things differently than I do, is it possible that he not only attaches different meanings to his words, but that his meanings are somehow basically inaccessible to someone like me? I will return later to the issue of incompatible perceptions, which has a much broader scope than mere creationism. Now Hume does what he can to support his contention that he simply sees things differently. For one thing he cites (p. 73) the somewhat indecisive supporting testimony of “ancient philosophers,” most likely Plato (p. 206). Many scientists working today have added their voices to the doubt whether design can actually be perceived at all in the physical universe (Scientific American, June 2009, p. 36 and Seed No 12 Nov./Dec. 2007 p. 20). Of course, when it comes to the issue of what are the facts, opinion does count for something. But by the same token we can be arbitrary in the weight we give to countervailing opinions. The range is from nearly universal agreement, such as the opinion that the sun will rise tomorrow, to a nearly even split, such as concerns many political issues. Where do we cross the line such that the weight of opinion contributes significantly to considering something an established fact? We do have some independent tests for judging the validity of an opinion. One is logical consistency. Do conclusions derived from observations interpreted one way violate any logical laws? If so, that is a very strong argument against the validity of the observations. Another is consistency with other theories that we hold to be true. This sort of inconsistency is nearly as strong a disqualifier as logical inconsistency. In fact it relies on logical inconsistency as an additional premise. Yet another is behavioral consequences. Does belief in certain observations motivate actions that we consider unacceptable, perhaps irrational? This is a weaker but still compelling test.

An insight we can draw from something implicit in Hume’s argument is that, if a theory accounting for the emergence of a phenomenon can be devised along the lines of incremental change, such as occurs in the organic realm, then the validity of that theory is a strong inclination to believe that the phenomenon is organic or at least very similar to organic entities. Philo does not state, but the alert reader can add, that, if there are signs of a history of incremental development in the apparently designed constituents of the universe, then we should be inclined to see the universe as an organism and not a product of “human artifice.” There are clear analogies to the theory of natural selection, where the issue is not the emergence or development of life as a whole but the emergence and development of particular species. Natural selection provides a mechanism for evolution and so provides a reason to believe that species evolve and are not the result of singular acts of creation. Some foresee the opportunity for a like theory of the evolution of the constituents of the physical universe such as would provide an equally strong reason to believe that the universe is not the product of a single act of creation. (The devising of such a theory may be a bit more difficult than Dawkins leads us to believe since it could very well involve an “evolutionary” explanation of current physical laws, a daunting prospect since it is hard to see how those laws would not be presupposed.)

11) The Concept of Design or Manufacture Based on Our Understanding of Human Design and Manufacture is Vague and Inadequate. (p. 67 and also p. 69, Hume's references to Cicero) There are two separate strands to this argument. The first is that the very complexity of the universe differentiates it from known products of human design and manufacture. Philo (p. 67 ff.) cites Lucretius and Cicero’s Epicurean as opinions that support his point of view. Hume adds that recent astronomical discoveries, far from adding evidence to the idea that the universe was constructed along a human model of construction, actually support the contradictory opinion. The idea is that designing and building the immensi summam and its “immense grandeur and magnificence” would require tanti operis, tanti muneris that we cannot even imagine any human or human-like artifice that is equal to the job. In one sense this isn’t Hume’s strongest argument. Since the designer is supposed to be “all-powerful,” it can supposedly overcome mere size restrictions. But Hume’s actual point is a bit different. His purpose is to drive a wedge between human design and construction and the supposedly analogous design and construction of the universe. Human design and construction is supposed to be the model for the fabrication of the universe. But because of Hume’s wedge, the analogy doesn’t work. All Philo wants and needs to do is to show the weakness of the design and fabricate model; he doesn’t need to prove the negative.  The second strand of Hume’s argument (p. 69) is that many products of artifice are produced not by a single great genius – some sort of incorporeal Howard Roark – but by teams of contributors or by decidedly inferior individuals who simply borrowed or pastiched the work of their predecessors. There is nothing in the appearance of a designed universe to exclude the possibility that the workshop of its creation was more like a Hollywood movie set or the office of some strip mall architect in the American hustings. There are, it turns out, some pretty serious limits on what a design argument, even if we were implausibly to assume its success, actually proves.

12) The Designing Mind Could Be the Universe Itself. Although Hume derives this possibility from the opinions of the “theists of antiquity,” it has an intriguing modernity about it, a sort of hippy-dippy appeal. More importantly it pierces the traditional objection that matter cannot think by regarding that assertion as based on one’s perspective. Imagine you were a teentsy tiny entity, about half the size of your average protein and that you could somehow observe, imagine, hope, reason, fear and, yes, even love. Imagine further that your domicile was a human body. What you think is the entire universe is actually the body of some schmuck living in Cleveland somewhere. You observe (using a teentsy tiny telescope) regularity and apparent purposiveness in the functioning of your universe. On a regular basis a million suns would explode in one region of the universe (the schmuck’s brain) followed by the appearance from out of nowhere of numberless organic molecules some of which would be split at unbelievable temperatures, and among those that weren’t turned into little merdes a few might bond nicely to you by some mysterious perfect fit. If you were of a teentsy tiny philosophical turn of mind you might think some outside agency, some external divine creator engineered this whole marvelous and purposive process. Little do you know that nothing external is needed. All the purposiveness, all the mysterious processes are the product of the universe itself. The schmuck is eating.

Notably, in the course of this argument Philo suggests the theory of mind most prevalent in modern science. Nothing, he says, was more repugnant to the ancients than the idea of mind without body, “a mere spiritual substance.” It is very possible that the mind “they felt,” was no more than body itself, described differently: “An order, arrangement, organization or internal machinery….”

13) Any Number of Cosmogonical Hypotheses Are Available As Alternatives to the Designer God Hypothesis. Particularly intriguing is the idea of an infinite number of universes actualized over the course of time (p.84). Intriguing not just because it calls to mind current multiverse proposals, but also because of its plausibility. Assuming the universe is infinite in some relevant way, and assuming that its contents undergo constant change, the result is an infinite number of configurations of the universe. It seems inevitable that one of these configurations should be the “designed” universe that we perceive. If the contents of the universe are infinite but exist all at the same time, then all possible configurations are realized simultaneously. If the contents of the universe are finite but it has existed forever, then all possible configurations are realized in sequence. Hume’s conclusion is certainly not necessarily true since a finite number of configurations might just be realized recursively an infinite number of times. However, Hume is not trying to prove the former alternative. It is enough to show that it is at least as plausible as the goddist hypothesis.

14) Motion, Change and Cause Need Not Be the Products of Human-Like Intentions. Or indeed of any external cause of any kind. This is, as far as I know, the first instance of a challenge to the time worn philosophical assumption that everything must have a cause, that something cannot come from nothing, that change requires an agent of change. That assumption was so pervasive that it formed the basis for the deistic proofs of the existence of god proposed by Locke, Voltaire and Hume’s own contemporary the pseudo La Mettrie. Hume begins by pointing out a number of phenomena, including “gravity, elasticity and electricity” that occur “without any known voluntary agent. (p.85)” To suppose such an agent is a “mere hypothesis.” An important distinction should be made here between the scientific imperative to look for explanations for phenomena and what Hume explicitly refers to as the hypothesis of a “voluntary agent.”  The former usually takes the form of a search for what we may call an immanent cause. If an object were to show signs of being attracted, our scientific imperative tells us to look for another object that is the cause of the attraction. Even the search for an explanation for the entire phenomenon of electromagnetic force takes the form of seeking to redefine that force in terms of other – hopefully simpler and more universally applicable concepts. Keeping explanations focused in that way is what distinguishes the scientific from the religious or animistic approach to explanation. And, of course, the appeal to an external humanoid but very possibly “perfect” voluntary agent is not only not an explanation, it closes off the search for explanations. It is the equivalent of saying, “Your research is at an end. Now go off and fast for forty days.” And, setting the scientific imperative aside, Hume is simply saying that events without causes are not inconceivable (i.e. their conception does not involve a logical contradiction). And he may very well have been the first to say so. (On p. 91 Cleanthes applies Hume’s point directly to the so-called causal proof for the existence of god that provided a great deal of intellectual titillation for 12th century theologians. Hume’s specific target, however, is deism and the aforementioned Lockean proof.) Furthermore it is compatible with the current view among scientists that certain phenomena such as quantum events and cognitively indeterminate behavior may simply have no causes. Incidentally Hume offhandedly remarks that the probability of past events is 1 (“…every event, after experience, is equally easy and intelligible….” This is an important way of saying that the calculus of probability cannot be applied to the facts and to past occurrences in particular. Philo’s simple “Why not” was a barb dipped in curare; it paralyzed the nervous system of Xtian and deist theology.

15) There Are No Examples in Experience of a Disembodied Mind Producing Material Effects. This is another formulation of the imperfect analogy family of arguments. The creationist wishes to form an analogy between the minds of artisans creating artifacts and the mind of some god creating the universe. The analogy, however, is seriously flawed. The mind of the human artisan does not create the house directly. That is to say, it does not have a direct causal effect on the building materials. It does have a direct causal effect on a material object, its own limbs, which in turn directly cause the transformation of the building materials into a house. The creationist model, on the other hand, asserts that god creates the universe much as if the artisan levitated the building materials by some sort of mental telepathy. Hume’s actual words are “…thought has no influence upon matter, except where that matter is so conjoined with it, as to have an equal reciprocal effect upon it. (p. 88)” Personally I don’t think this is the happiest formulation. The idea of reciprocal effect is confusing and perhaps unnecessary. What Hume could say and perhaps wanted to say was that, in the known cases of the influence of “thought” on matter, thought so-called was in fact identical with a bit of matter. The brain and the nervous system of the artisan is his mind, which has a direct causal relation to his muscles and so on. Another breakdown of the analogy Hume does not mention is that no human artisan creates a house ex nihilo. He needs materials from which to work. The creationist idea has god operating in some big empty black space.

16) An Infinitely Long Chain of Events Cannot Have a First Cause. Obviously the concepts of an infinitely long series and of a first element of a series are mutually contradictory. Stated this way, it is not strongest argument since not all religions regard the universe as having been in existence for an infinitely long period of time. However, it is worth keeping in mind since the point is made (pp. 92-93) in response to Demea’s concession that the chain of events in the universe may extend infinitely into the past (He goes on the argue that the entire chain still needs some cause standing outside of it).

Parts X-XII of the Dialogues comprise a discussion of the problem of evil. This is not so much a refutation of the design argument as a very powerful argument against the supposed existence of an omnibenevolent god. Hume does take time to remind us (p. 107) that his arguments concerning design are answers to the supposed proofs of the existence of a designer god. They are not themselves disproofs of that god’s existence. The observable traits of the universe may be consistent with the universe having been designed; that does not mean that there is anything about the universe that forces us to conclude that it was designed. It is worth giving some historical context to the issues of design that are raised in the Dialogues. By the beginning of the 18th century Maxju theology was reeling from the brutally insightful scriptural criticism of Spinoza and Bolingbroke. No rational observer could take the Bible and the Old Testament in particular as the basis for any acceptable religious belief. In place of Xtianity European intellectuals developed the belief system we call deism (When Hume says “natural religion,” he by and large means deism just as Jefferson signaled he was a deist and not a Xtian when he spoke of “nature’s god”). The roots of deism lie in Spinoza’s tantalizing if elusive identification of god with the universe or nature, as well as in Locke’s fairly abstract discussions of god in the Essay Concerning Human Understanding (especially the first edition). Pope and Voltaire turned these ideas into something of a system. The idea was to define god as no more than an abstract force that created the universe and (in some versions) keeps it in existence. Two proofs were proposed of the existence of the deist god. The first, endorsed and restated in different versions by Descartes and Locke, used the concept of cause. But causal proofs, or indeed any sort of proof that was in any way abstract and did not have a firm grounding in experience, died a slow death after Locke. Historically speaking the conviction was gone, and no philosopher of any merit seemed able to muster the energy to defend the causal proof.

This context gives us some ground for an alternative reading of Hume’s Inquiry and the first part of the Treatise. Yes, Hume provides a conceptual framework for understanding the inductive scientific method. And yes, he also demonstrates that many conclusions that had appeared to be a priori, or, as the saying went, geometrical, were in fact extrapolations from often repeated singular experiences. They were generalizations and not deductions. But it is equally important to note that Hume’s assertion that our very idea of a cause is just the product of somewhat arbitrary associations based on experience and consequently that it had no validity of its own outside our associations – this assertion robbed any sort of conceptual foundation from the causal proof. If you apply what Hume says to causal proofs from Aristotle to Locke, you find that the proofs fail because there is no meaningful concept of cause on which they can be based. The idea of cause itself (or habitually observed succession) is faulty, something of an empty abstraction. Now, assuming that the causal proof failed or that it simply lost steam in the 18th century, we have some ground to conclude that goddists began to appeal to the same experiential evidence that formed the backbone of empiricism. The new assertion was that god’s existence was proved by the evidence of the senses. Few really dared assert that they “saw” god, perhaps fearing the same scorn that greeted the Jansenists. However, it was acceptable to assert that, in observing the universe in exactly the same way natural scientists do, they “saw” that it was obviously created by something. It appeared created because it showed signs of workmanship or design. This, I opine, is the origin of the design argument. It is a nefarious attempt to hijack the empirical spirit for goddist purposes. In began to show up in different forms in the writings of Voltaire and his opponents and is strongly stated by the pseudo-La Mettrie as well as numberless now happily forgotten 18th and 19th century theologians. Paley’s restatement of the design argument in popular terms is the notorious watch in the meadow paradigm. Unfortunately the cowering members of the post-Napoleonic generation forgot Hume’s forceful logical argument. It took the theory of natural selection and the important distinctions between order, design and creation to put Paleyism out of its misery. Paleyism still has its adherents particularly among the less literate inhabitants of the American backwoods, and regrettably the prominence of the American economy and military put us in a position where we have to deal with these creatures. Fortunately, since the collapse of religion among the educated and in Europe generally around 1855, creationism has ended its life as an historical curiosity unrelated to serious science or philosophy.

So much for Hume. I have not found in the great empiricist any trace of what I consider to be the strongest argument against anyone who disingenuously asserts that the entire universe looks designed to him.

II. You Cannot Use Examples To Show That The Entire Universe Is Designed (Even If Your Example Is The Whole Universe)

Hume’s refutations of the argument from design are largely, as I said above, based on his own empirical (i.e. based on observation) theory of analogy or induction. They come down to the insight that the supposed evidence for design, the perceived “designedness” of the universe is itself an analogy and a very weak analogy to boot. However, there is a much more serious logical flaw (The logic in this case is best expressed in the theory of complementary sets. There is, however, an epistemic element to the fallacy. Complementary sets may be so defined that one of the complements is unknowable. If that is the case, nothing, aside from axiomatic deductions within a system, can be successfully proved about the unknowable complement.) in any argument based on a perception of the universe as designed. The concept of a designed universe bears a remarkable kinship to Hegel’s initial concept of being or to the set of all non-self-containing sets . All three run into logical trouble. While the idea that the entire physical universe is designed is not completely self-contradictory, such as a purely logical concept might be, nevertheless it is such that no possible example could be produced to show that the entire physical universe is designed, not even the universe itself. For the apparently designed item can only be apprehended as apparently designed if it can be meaningfully contrasted with another perceivable item that looks undesigned. (George H. Smith, pp. 267-268,  proposed this very argument using physical and epistemological language: We identify objects in the universe as designed or undesigned because we possess the concept of design and its complement, natural. We possess these concepts because we have learned to recognize characteristics of design in the objects around us. The universe as a whole, which is the very paradigm of what it means to possess natural characteristics, cannot be described as designed because there is nothing "really natural" to contrast it with.) However, if the apparently undesigned item were in fact designed, then there would be no cognitive or evidentiary contrast with the apparently designed item. The latter would not be distinguishable from the apparently undesigned item and its appearance of design would be an illusion. This is an evidential or epistemic contradiction and not a purely logical or conceptual contradiction (Cf. Logical Counterparts). It comes down to the fact that if you use a particular instance as proof of a general statement about things that are not relevantly similar to the particular instance, then you deny an important fact about the particular instance that you had wanted to assert and indeed highlight. It comes down to saying that anything that is F is also G, that you have discovered some F and thereby proved that not-F’s are also G. Aside from being a false analogy, as Hume observed, the conclusion of this argument to the effect that all non-F’s are G’s also undermines the value as evidence of the initial assumption that if something is F then it is G. The conclusion just makes the initial assumption an instantiation of some general assumption that everything is G. But of course it is the general assumption that is in question.

The creationist has two possible approaches. Either he can say that part of the universe looks designed; therefore the entire universe is designed. Or he can simply say that the whole universe looks designed to him. Neither approach passes the epistemological sniff test.

I like to fancy that this argument is superior to any of Hume’s formulations, not just because it points out an epistemic fault (the fault lies in the faulty way certain statements are used as evidence for other statements) and so it doesn’t just disincline us toward but absolutely flattens design theory absurdities (Hume’s bouquet of anti analogy arguments might be compared to carpet bombing whereas an argument that shows logical inanity is the hydrogen bomb). It undermines the foundation of the creationist dogma - its evidence. Creationists say they “see” design in the universe. Hume responds that, whatever they may see, it is disanalogous to what we understand as design. The argument to logical or rather epistemic fallacy shows that, once creationists conclude that apparently undesigned things are really designed, they never really “saw” design in the first place. Alternatively if they see everything as designed, they dilute the term “design” of any meaning due to lack of a meaningful contrast. (Creationists should at this point be cautioned against the old ineffability song and dance so beloved of the local priest. If the whole business is so fucking ineffable, you really didn’t have to bring up your “evidence” in the first place, did you?) Looking designed on the part of things that look designed is not a real distinguishing characteristic in the creationist conceptual framework, since it is possible to look undesigned while being designed. The design proof doesn’t add anything to the statement that it is logically possible that anything could be designed. But we don’t know which is which since there are no grounds for recognizing the distinction in reality.

The historical circumstances that led to the devising of creationist design theories have no more than a kind of antiquarian interest. But they do shed light on a real philosophical pickle that Xtian theologians find themselves in. By the early 18th century in Western Europe scholastic “proofs” of the existence of god had lost a great deal of their credibility. The causal proof was still on the table. However, one important effect of Hume’s Treatise and first Inquiry was to show that the concept of causality was empirical and not logical and by implication that it was unsuited for purely logical proofs without verification (Show me the first cause). What we know as the design theory started life as a deist slogan. Yet its apparent intuitive plausibility gave it over time the appearance of an actual proof. There is actually a stump of the design slogan in Plato's Laws (Bk X 886A and Bk XII 966D-E ) between glowing descriptions of proto-fascist youth camps and proposals to castrate the poets. The real action follows with the fatuous self-mover proof in the same Book that the combined efforts of Spinoza and Hume would turn to mush. Hume showed that design proofs aren’t actual proofs of anything. I hope to have added that design is in fact a logical shambles. The result is that there is no real Xtian theology left, nothing that distinguishes it from, say, scientology.

To be fair, not all the examples cited by creationists are presented with the sole intent of proving that because one thing is apparently designed, then other things that are apparently undesigned are in fact designed. Some examples are simply proposed as phenomena that alternative explanations cannot account for. You might call this creationism lite.

Creationism lite, however, is not less filling, and it really tastes like shit. A charitable reading of the now discredited creationist claim that certain bacteria like E. coli are far too complex (The phrase is “irreducibly complex”) to be explained by mutation and natural selection (Charitable only if irreducible complexity is not understood as not distinguishing in fact or in epistemic status – our ability to distinguish two things that may be distinct in reality – biological entities are that are irreducibly complex from those that are reducibly complex; if that distinction is made, then, if the createdness of the irreducibly complex is presented as evidence for the createdness of the reducibly complex, our ability to distinguish between the created and the non-created vanishes; irreducible complexity can no longer function as a criterion for distinguishing the created from the non-created) would be that it just picks out certain phenomena for which there is no satisfactory scientific explanation. According to this reading, Behe’s claim is an example of creationism lite. It merely says there are some things natural selection cannot explain; it does not or should not make any stronger claims. To say that one explanation for a natural phenomenon fails is not equivalent to saying that the alternative, namely creationism, is true. But creationism lite really doesn’t say much of anything. At best it amounts to the injunction to keep looking. It certainly doesn’t say what the creationists want it to say, namely that the only possible explanation of some properties of E. coli is that an entity possessed of intentions created it. Nor does it say that the conclusion that E.coli was created entail that biological entities whose properties can be explained by natural selection are also designed and created -  a conclusion that, as I indicated above, leads to epistemic suicide. (For another approach to answering the claims of creationism lite, cf. Paulos, pp. 20 ff.)

John Leslie (whose talents as a philosopher rival my own abilities as an NFL quarterback) is a different kettle of fish. His idea is that if we saw “God made this” scribbled on a mountainside somewhere and if we had good reason to believe that this was not just those crazy guys from the Templeton Foundation up to their usual hijinks, then by a series of “obvious” intermediate steps, we should conclude that the old moving hand not only wrote the inscription, but also that innocent bystanders like the bits of shale surrounding the inscription were in fact designed and created by the same moving hand. With this conclusion Leslie throws himself under the epistemological locomotive. If you obliterate the relevant distinction between the inscription and the innocent bystander (i.e. that one is designed and the other isn’t), then the fact that the inscription appears to express an English language proposition is not sufficient to distinguish it as having been designed or created. We need not even go into to alternative possibilities like irregular mineral formations into which Leslie projects a personal gestalt, or that the inscription could be the work of a half-assed designer god who designed the inscription but not much else. The Great Designer could also have just put the inscription there in one of its more cutesy moods as a way of “testing” us mere mortals (Of course the things that look designed might be clever clues on the part of some shithead god like the pope in the pizza, a sort of secret language of the flowers to keep us guessing). Either way it’s Leslie who has assumed the burden of proof and his proof fails.

Paley's infamous watch is pretty much in the same boat. It depends on a self-contradictory epistemic distinction between things that look designed and those that don't look designed but are designed, leading to evidentiary futility. There is an obvious logical fallacy in the Paley watch argument. We learn by experience (ostension?) that certain things are created from raw materials by human craftsmen and we come to recognize by a family similarity other things that belong to the same category. At some point, whenever we see something that has sufficient observable similarity to our original examples, we make the assumption that those are the same kind of thing (Watches, for example); further, we conclude that, since Bulova made the examples we learned about in First Grade, then Bulova or some equivalent to Bulova (and not Seiko or Citizen) must have made new examples that we recognize as the same kind of thing. The Paley example starts with an assumption not unlike someone who concludes that any old watch found in a field must have been made by Bulova and only Bulova. This is bad enough. But there is more. We recognize watches because we can contrast them with other things. If not we wouldn’t be able to distinguish watches from toasters. In the same way we recognize things as manufactured only if we can contrast them with other things not manufactured. Otherwise we wouldn’t be able to distinguish a genuine Bulova from a chunk of meteorite. Paley simultaneously distinguishes the watch from the stone and asserts that in the relevant respects the watch and the stone are indistinguishable. If everything is artificial, nothing is artificial. I think a good case could be made that Paleyites and their creationist successors suffer from visual agnosia. They commit a fuckosophical fallacy we may cheerfully call the Agnosiac Fallacy.

Take a closer look at how we identify a watch as artificial. We approach it with certain assumptions. We should know what a watch is, or, failing that (I have a feeling that this watch example is going to need a lot of historical footnoting after a couple of generations), we should recognize that certain metals are found nowhere else in a natural state. We recognize the numbers and understand that certain sharp edges and perfect figures have always in our experience been the product of machine tooling. These underlying assumptions and background experience do not apply to our apprehension of smiling meadows not to speak of the universe as a whole.

Of course some versions of design theory (Is “theory” perhaps too grand a term? Doesn’t it put this nonsense in the same class with Newton and Darwin?) don’t rely on examples. Some people just look up at the starry sky, get this kind of dreamy faraway expression on their faces and whisper, “Don’t tell me that somebody didn’t make that.” There are in fact two distinct creationist assertions, both of which succumb to the Agnosiac Fallacy and another fallacy that we might call the Universalizing Fallacy. The universalizing fallacy consists in an improper inference to the effect that, if one thing is F, then everything is F. The agnosiac fallacy consists in the improper inference to the effect that, if one thing appears to be F then things that do not appear to be F are also F. The first assertion is that everything is designed. The second, the form favored by star gazers, is that everything appears designed. The second, being a much stronger claim, is actually much more difficult to sustain. If everything appears designed, then it is impossible to point out anything that does not appear designed. There is no criterion for something to appear undesigned and therefore there is no working criterion for something to be undesigned. But if we cannot specify or describe what it is for something not to be the case, we cannot really describe what it is for it to be the case. Someone who asserts that absolutely everything we do is sinful forfeits thereby the ability to describe what is not sinful and so forfeits any meaning at all for the term “sinful.”  Someone who says absolutely everything he sees and, by implication, you see appears to be green forfeits thereby his ability to describe what is not green. For he would describe what we might now consider red as really green. Making assertions involves drawing boundaries; universalizing assertions of the kind that everything appears to be designed wipe away the boundaries. If absolutely everything looks to farmer Jones like it belongs to farmer Jones including those plots that had been fenced off by farmer Smith, then farmer Jones has no comprehensible concept of property. The first assertion seems stronger, since the argument might be made that everything is designed, but only some things look designed. Those things are hints placed under rocks by the clever designer to suggest its presence. But the concept of being designed thus used is just as meaningless or rather epistemically vapid. For in every case it breaks the link between appearing designed and being designed. If everything that appears undesigned is actually designed, then we have no way of describing what could in fact be undesigned (and eo ipso designed). The alternative cannot even be communicated. If every possible configuration of little Johnny’s room is in fact an intention of order on Johnny’s part, despite its apparent messiness, simply because we find his collection of Playboys in a drawer to be arranged in a possible succession (by date perhaps or centerfold boob size?), then we cannot communicate what it would be for Johnny’s room to actually be unintentionally disordered even with respect to those parts of the room that appear to be a mess. (It is, of course, possible to embed the concept of design in indirect discourse. Something on the order of: “That thing looks natural (undesigned), but I assert that it is designed.” Indirect discourse in a way shields the embedded predicate from having to be meaningful to whomever is addressed. All the addressee needs to know, at one level, is that the speaker of the sentence, asserts something. But if a creationist opts to shield his meaning in this way, he also loses his ability to communicate, much less prove, what he means when he says that anything whatsoever is designed. Hiding meaning behind “that” is exactly that – hiding it.) Both versions end up asserting no more than, “Everything is as it is,” an assertion to which we may heartily subscribe. We would only add, “It is not the case that everything is designed.”

It is also worth noting that conceptual distinctions and epistemic distinctions do not imply that objects corresponding to what is distinguished actually exist in reality. Every predicate divides the world into two groups, those that have the quality denoted by the predicate and those that don’t. But it is not the case that the existence of entities characterized by a predicate requires the existence of entities characterized by the complement of that predicate (i.e. that the proposition that the former exist entails the proposition that the latter exist). The quality of conformity to the inverse square law as it applies to mutual attraction between physical objects has as its complement the quality of not conforming to the inverse square law. However that is not a reason that there should not be a physical universe the attraction of all of whose physical objects conform to the inverse square law. Our own appears to be an example of one such universe. To draw conclusions about things that exist on the basis of logical distinctions is a notorious scholastic fallacy. The “logical counterpart” of this fallacy is often termed “empiricism.” (Is there something wrong with the fact that I speak of things existing here even though I agree with the stricture that existence is not a predicate? Not in my opinion, but a complete explanation of why that is so would involve a wild digression. Suffice it to say that when the English expressions “exists” and “does not exist” are meaningfully and legitimately predicated of an entity, “exists” is used as a shorthand for real predicates whose meaning can be specified without contradiction.) Why do I make this point? Read the next paragraph.

The functional value (some would call it the epistemological value) of a conceptual distinction appears when that distinction plays a role in a proof about the world. The design theorist seems to profit from Mackie’s argument in that one could conclude that, while there must be a logical distinction between the designed and the undesigned for the concept of designedness to make any sense, still our universe might be entirely composed of things that have been designed. Not so. The design proof purports to be a proof by showing us examples of things that have been designed (or at least look designed) but have not been designed by humans. If, however, everything is in fact designed, then there is no real or truly observable difference between things that look designed and things that don’t look designed. The basis of the proof, the supposed perceivable difference between the designed and the undesigned, evaporates. To use technical jargon, there may be a logical distinction between the designed and the undesigned, but there is no functional or epistemological distinction. The claim to see things that appear to have been designed by an unknown hand turns out to be false. (If we set the design argument aside and choose to work with purely ontological parameters, then it may indeed be the case that apparently undesigned things were in fact designed. I mean that the concept of design might be genetic and not descriptive. A thing is designed because someone designed it and not because of any observable qualities. There is no “essence” of designedness that we can point to outside of someone stepping up and claiming the credit. However, what is at issue here is the observational argument that things are designed because they look designed, an argument that remains invalid whether or not it is genetically true that most things, or indeed everything, were in fact designed.)

Let’s return to our color example. The logical complement of green is not-green. The concept of greenness makes no sense unless there is something like non-greenness. Now let us throw a St Patrick’s Day party where only people dressed entirely in green would be admitted. Everything else was also painted green: walls, ceilings, skin, hair - everything. In the room in which that party was held non-greenness does not exist. Obviously this situation is not logically impossible even though non-greenness is a logical complement to greenness. Now let us assume one of the party goers had been blind from birth. He suddenly recovers his sight at the party. A mischievous partygoer tries to convince the formerly blind man that everything outside the room is green also (Quite obviously the formerly blind man has no real idea of what he is talking about – and, past a couple of extra steps, that is exactly the point). Meanwhile a bit of paint has rubbed off one of the walls and shows white underneath. The mischievous partygoer points to both a green patch of wall and the white patch of wall and says that they are both green. If, according to the mischievous partygoer, everything in the world is green even if it is what we would call white, that is, if he defines everything as green whatever the color, then his proof that everything is green fails as proof. It is functionally and epistemologically useless in proving whatever it is he wanted to prove (Perhaps an inference that, since everything is green, St Patrick likes mints). It is in fact just a stipulative definition. In the same way the design argument that everything is designed is merely a stipulative definition and useless in proving the existence of a designer. I could as easily stipulate that I define everything in the world as made by Toyota because some things in the world look like they were made by Toyota (on the basis of the old definition of “made by Toyota”), even those things which, by the old definition of “made by Toyota” do not appear to be made by Toyota.

How do we distinguish Stonehenge or Easter Island from natural rock formations? Or Mount Rushmore from the Old Man in the Mountain in New Hampshire? If the Old Man in the Mountain was also a sculptor’s design, then the impressionistic evidence disappears for distinguishing that rock formation from Mount Rushmore. As does the reality of the distinction. To say that every rock fall was sculpted because Mount Rushmore looks sculpted is to empty the concept of being sculpted from any meaning whatsoever. The distinction based on visual impressions may in some cases be difficult to make (This is a problem for the style history of art as well). In that case independent factual evidence distinct from otherwise unsupported impressions may be our only recourse. It may be that the only reliable basis for distinguishing North American Indian burial mounds from natural formations is the presence of artifacts inside the mounds. Proof may be lacking without some sort of definite ontogenesis. This mirrors biological classification where phylogenetic evidence always overrides arbitrary associations based on resemblance. To relate to the issue at hand, you still need external proof that there was an actual designer to support your visual impressions that certain things look designed. But if you still need external, presumably ontogenetic proof, then the “design proof” was never a proof in the first place.

A couple of generalizations are in order here. There is a distinction to be made between creation and design. The latter is a stronger concept. A mother creates her child but did not design him. If planets were ever formed by material thrown off by a star, you could say the star created but did not design those planets. Design involves intentions. Since design is a stronger concept, proof of design (or rather activated design, since a mere design may languish on the drawing board) is, of course a proof of creation. But proof of design might be a bit more elusive than simple gap toothed bewilderment at a chrysalis. We should also distinguish between proof of design and proof by observation. What Paleyites do is to use a supposed proof by observation to produce a proof of design. Proof by observation can be strong. Understood broadly, observations lie at the basis of all proofs (including logical proofs in my humble opinion). However, proof by observation alone of what we may, bastardizing Locke, call a secondary quality is a very narrow understanding of observation and is considerably weaker than observational assertions in a broad sense. There are two sources of the weakness. The first is that secondary qualities are notoriously difficult to universalize. How can you show a dumb jock that his velvet painting of poker playing dogs or Michael Fried that his treasured photograph of trees is not the greatest thing since sliced bread? The second is that singular unsupported observations can most often be trumped by networks of mutually supporting observations, such as when phylogenetic evidence trumps similarity based classifications.

Any example of design that we know is an imposition of form on matter. The claim that everything is designed, even the ὕλη itself so to speak, does not use any definition of design that we can understand. It seems to me that design presupposes a state change. Something unformed is given form and sometimes purpose, and we call the prior state undesigned and the latter designed. I further believe that it would be difficult to reconcile this intuitive understanding of design with the Xtian doctrine of creation ex nihilo. This argument is a weaker version of the epistemic contradiction argument I outlined above, since it is not an outright contradiction to assert that a designed entity appeared that had not been created from some unformed matter. Nor do creationists need to assert that designed entities did in fact appear without the pre-existence of unformed matter. The existence of mysteriously designed entities is just evidence for a process whose details don’t have to be spelled out. For example, god could have created the fully designed entities ex nihilo, much like casinos appear overnight on what was waste land. Or else god could have created the “unformed” matter first and then molded it into designed entities. Whatever. The argument is not conclusive, but I do find it persuasive.

Note the creationist is not just trying to prove that it is logically possible that everything is designed even though some things appear undesigned. He is trying to convince us, on the basis of a few things that appear designed despite expectations, that everything else is designed also. But if he is to convince us his conclusions must be stronger than sheer logical possibility. After all it is also logically possible that everything is undesigned. The designedness of everything might be compatible with some things appearing to be undesigned, but this is useless as long as we are still at the stage of trying to infer the designedness of everything from the apparent designedness of a few things. The fact that something is logically possible does nothing to support the case that it really obtains.

The Agnosiac Fallacy that we unearthed in our investigation of creationism is a logical fallacy, or, more accurately put, a conceptual error. The conclusion of the argument (Everything is designed) contradicts a necessary assumption of the argument (I can reliably identify some things as designed because they look designed). Note accusing some of committing the Agnosiac Fallacy does not involve any claims about the facts. That is, everything may in fact be designed even though only some things look designed. The accusation merely targets the validity of the argument, viz the argument to the effect that the apparent designedness of some things proves the actual designedness of everything else. But we’ve gained some useful material above and beyond the limitations of the creationist fallacy. We have found that a report of an apparent “obvious” perception of a secondary quality can be evaluated for logical errors. Sometimes it takes work to tease out the error. Paley’s mistake was lurking somewhere behind the mainspring of his celebrated watch. (Side note: Schlick (pp. 133-134; cf. also p. 177), citing Hobbes, Mill (pp.64-65) and Bain, employs a version of anti-agnosiac reasoning to establish his point that a unified consciousness must be able to experience differences. The Agnosiac Fallacy violates Bain’s Law of Relativity.)

Side note: Without going to the complex and not immediately relevant trouble of defining exactly what I mean by "dialectical," let me just observe that this argument for the failure of evidence for certain universalizing assertions, or for the self-defeating nature of behavior such that we may reasonably conclude that anyone engaging in such behavior believes in the truth of those assertions, is a dialectical argument. That may be a reason why empirically minded philosophers (e.g. Mackie) don’t recognize it when it’s waving a red flag in front of their collective noses. Philosophers of a Marxist turn of mind or background (Sartre or Heath & Potter spring immediately to mind) are most at home with dialectical arguments. Husserl, in the midst of an orgy of dogmatizing, dipped his toes in a dialectical argument (Cf. Ideen, p. 157, Ftn 2). The form of his argument resembles those surrounding negative existentials. Even Berlin uses this form of argument with a certain ease.

I shall conclude this section with the simple remark that creationism is the weird twin of the argument from miracles. One says the universe is not uniform. Therefore god exists. The other says the universe is uniform. Therefore god exists. Of course, like the Inquisition, the goddists have already made up their minds what the conclusion will be, so they use whatever facts come to hand as their premise. Or rather they’re like the superstitious old lady who attributes all deaths to murderous leprechauns. Her explanation unfortunately is even less likely than a spontaneous occurrence. Equally unfortunately, if it is false that god exists, then that falsehood could only be truthfully implied by another falsehood (a fact that could go a long way toward explaining the cavalier Xtian attitude to the truth).

I should imagine the sophisticated fuckosopher already suspects there are deeper issues lurking here than what we may have found in the carnival of creationism. And they relate to the nature of observation and the role observation plays in philosophical and scientific proofs. Accordingly:

III. Proof by Observation

It is not uncommonly held that logic and observation (in a broad sense of “observation” that includes collecting data in the framework of commonly understood concepts and classifications as well as what Plato rather grandly called apophansis, viz the discovery or invention of a concept or some sort of thing in the world. By the way, Aristotle dumbed down this important insight to something that his translators render as “proposition”) constitute the twin bases upon which philosophically valid conclusions can be constructed. Logicists, however, sometimes lose sight of the unruly observational component of proof.

Let’s start by pulling up a few instances where the validity of an observation is not just a contributing factor but the key player in drawing a philosophical, scientific or other sort of conclusion. Obviously the gorping version of design theory relies entirely on a supposedly undeniable observation: The universe just looks designed to me. Even the version that says some portion of the universe to which we cannot assign a human cause looks designed combines the argument that there is no human cause with a supposedly valid observation. The Behe version, however, does not rely on observation alone, and – This is a significant difference from garden variety creationism – Behe does not rely on evidence from observations that anti-creationists (often dubbed “scientists”) do not share. His argument is not that some portions of the universe “look” designed. Rather he contends there is no current explanation for some phenomena other than intentional design. Swedenborg’s encounters with angels fall squarely in the otherwise unsupported bag of observations. (An interesting exercise would be to distinguish Swedenborg’s observations from Rilke’s.)

There is no sharp distinction between Maria’s vision of Pope Armando and the creationist’s claim to “see” traces of design in the physical universe. One telling point of similarity is that both refuse (do refuse in the case of goddists, and would refuse in the case of Maria) to accept argument to the effect that there may be something wrong with their eyesight (Cf. Paulos, p. 19) Lacking such a distinction, there is no basis for treating the cases differently. There is no reason why the acceptance or rejection of the one should not entail the rejection or acceptance of the other.

But arguments based exclusively or largely on observations or perceptions that others may not have now or previously extend far beyond the realm of crackpot theology. Lacking a clear distinction or criterion for distinction, there is no evident reason for treating the observations of Maria or crackpot theologians any differently from the observations of natural scientists. Biological classification based on observed similarities assumes shared perceptions of those similarities. To a degree all empirical science is based largely on observation – though, as a science develops, an existing and accepted body of knowledge, mathematicization and perhaps contextual definition come to play important roles in situating and tempering new observations. Even the truths of formal disciplines like mathematics and the formal generalizations of the empirical sciences can be described in terms of a type of observation. The object theorist Alexius von Meinong, quoting Mach (p. 60), talks about an “immediate insight” (Unmittelbare Einsicht) we have to the effect that the laws of both logic and physics obtain – are valid. (Note “law” means one thing in logic and something else entirely in the natural sciences. Logical laws are axioms and rules of inference for a language. Physical laws are statements of observed regularities in nature.) The humanities and human sciences are particularly dependent on novel observations that the scholar tries to get others to share. The history of art as a history of style is in large part a classification of different types of perceived pictorial space – and the art historian must induce the reader to “see” different spaces or spatial configurations in works of art in order to successfully argue an historical point (developmental sequence, dating etc.).

Wittgenstein’s famous discussion of seeing-as includes aesthetic apprehension as an example (p. 173). Needless to say Wittgenstein was not concerned with the issue here of legitimate vs. illegitimate perceptions. He was addressing the phenomenon of a single object and a single sensation resulting in two equally legitimate seeings-as.

Literary hermeneutics is a technique for inducing the reader to perceive something in a literary work which was not obvious on a first reading. The philosophical method called phenomenology is almost entirely dependent on introspective observation. But the reliance on specialized observation is not unique to phenomenology. Sense data theory not only situates veridical observations (sense data) within scientific method, it also relies almost exclusively on an unjustified observation: the fact that we apprehend sense data at all. So-called ordinary language philosophers regularly assert without justification that “we” use words in a certain way. Specific philosophical schools aside, I suggest you count the number of times in a philosophical text that unsupported observations are made about language, the world or our mental make-up (They are easy to miss because these observations will seem the most obvious to you and without need of justification). A terrific example is Aristotle’s Categories, still along with the rest of the Organon largely the basis of logic. Aristotle makes his grammatical and ontological distinctions and gives a few examples and simply relies on your “seeing” that these categories are valid. The foundation of logic, so to speak, relies completely on argument by unaided observation (observation of language in this case and only indirectly of the world).

There is nothing wrong with argument by observation. In fact we couldn’t get along without it (Note, lacking some sort of empirical data, even the previous sentence could be construed as an otherwise unsupported observation. Indeed much of what I am going to say in the rest of this paragraph is otherwise unsupported observation. Wittgenstein (No. 485), truistically echoing Aristotle and Hume, observed that justification by experience has to end somewhere or else it’s not really a justification. Presumably an infinite series of grounds doesn’t justify anything). However, its very importance indicates it merits more attention than it usually gets. The validity of our observations lies at the very basis of empiricism and the scientific method. For, if we cannot even agree about what constitutes a basic observation, then we cannot begin to draw conclusions about our observations. Observations may be regarded as constituting in and of themselves a proof (of something). Or else proofs may be constructed using observation as a starting point. Anyone who insists that the 24-inch thin crust shows all the features of Pope Armando XVII to him needs to be reminded that he is offering us a proof (of his own observation report if nothing else).

It should be just as clear that we often don’t agree about what we observe. In fact that happens all the time. What might be right in front of your eyes might not be right in front of mine or anywhere in my field of vision. It is important to remember that we quite often don't "see" the same things, that seeing as a mental phenomenon is conditional in the sense that it probably never is pure, that it is always influenced by some factor or other.

We need to take note of situations where reliable observers report that they don't see something or that they see something differently from other reliable observers. Literature is rife with such situations. Take Balzac’s Le Chef-d’oeuvre inconnu where one man "sees" qualities in something (in this case artistic perfection in his painting) that others cannot. One literary critic has observed that figurative language is not "natural." We must come to "see" poetic imagery through a process partly conditioned by already existing mental schemata.

Toute relation de deux (ou plusieurs) mots coprésents peut donc devenir figure; mais ce virtuel ne se réalise qu'à partir du moment où le récepteur du discours perçoit la figure (puisqu'elle n'est rien d'autre que le discours perçu en tant que tel). Cette perception sera assurée soit par le recours à des schémas bien présents dans notre esprit (d'où la fréquence des figures fondées sur la répetition, la symétrie, l'opposition), soit par une insistence particulière dans la mise en évidence de certaines relations verbales.... (pp. 41-42)

Clearly these are examples from what I shall call second order perceptions. But conditional seeing also occurs at the primary or supposedly unconditional level, even in science. Freud noted exactly this phenomenon in a passing comment on how we must be "taught" to see through a microscope.

Auch der Student, der die ersten Male ins Mikroskop guckt, wird vom Lehrer unterrichtet, was er sehen soll, sonst sieht er es überhaupt nicht, obwohl es da und sichtbar ist. (vol. XI, p. 454)

Veridical or trustworthy observation and what constitutes it is an open question in the hardest of empirical sciences. Data discovery and hypothesis testing are obviously observation dependent. Mach was of the opinion that thought experiments also involved an element of observation (Cf. Meinong’s discussion of das Gedankenexperiment pp. 67-77. It goes without saying that Meinong’s Einsicht is not a viable solution to the indeterminacies of empirical observation.)

Closely relevant to the issue at hand is O'Hair's assertion that she simply failed to see harmony (i.e. design) in the universe (pp. 147 ff.). Lest we be tempted to dismiss O'Hair on ad hominem grounds, we should remember that many and probably most scientists consider the perception of harmony or design as absurd. Indeed an indirect proof of the implausibility of direct perception of design or harmony or whatever is the felt need to construct logical proofs of the existence of god. Those from Aristotle onwards who contrived logical proofs of the existence of god as well as all the negative theologians often asserted or at least conceded that they had no direct observational experience of god. If they had, the proofs and the negativity would have been unnecessary. After all Aristotle never tried to prove the existence of Plato.

So if there is no universally shared experience of what is seen, where does this leave the justification of beliefs which may be the product of observations combined with arguments or of observations alone? If the justification for a belief is an observation without an argument, then, if I am to accept that belief, I must make the same observation, i.e. have the same perception. Or, failing that, I must give more credence to the observer who made that observation as a witness than I would to arguments against his belief (This is Hume’s point). Since, by definition, a non-standard perception is incapable of the first sort of corroboration, it can only be justified by the second kind. Otherwise the belief, like the observation is private and unshared. Less charitably put, the observation is an hallucination and the belief a myth.

 A couple of distinctions need to be made. The first is between shared observations and what I shall call “alternative observations,” viz. those observations that can be challenged with some degree of justification and often are challenged. (The advocates of the pope in the pizza propose an alternative observation. Maria needs to describe her vision in such a way that we see the same thing, such that we “see” not only the pattern, but also its supernaturally intentional placement on the pizza. Failing that, her observation is subject to justified challenge.) Alternative perceptions rely on shared perceptions. They can’t get along without them. Without shared perceptions we couldn’t know that someone had an unshared perception. At the very least the proponent of an alternative perception wants to communicate his perception with us and that implies a shared language, a shared perception of his utterances, an understanding of his words and a common fund of memories and associations on which he can base his claim for a new perception.

Other distinctions relate to the challenge itself. Obviously anything can be challenged. All you need to say is, “I don’t think so” or “Not to me it ain’t.” Accordingly we should note the difference between Pyrrhonic skepticism and working disagreement. Pyrrhonic skepticism is not at issue here, although it could be addressed in its own time and place. Alternative observations are not alternative because a Pyrrhonist doubts everything. Alternative observations are observations or sets of observations that have in fact been challenged against a background of an overwhelming number of shared observations (“Overwhelming number” as I am using it is an intuitive notion unresponsive to mathematical fiddling. We may disagree with the observation that the universe looks designed and yet agree that stars look like points of light, that there are nine planets and so on. The points of agreement overwhelmingly outnumber the point of disagreement. Of course you could make the number at least equal by appending, “…and that looks designed” to every point of agreement. But that is what I call mathematical fiddling.)  Along with the distinction between Pyrrhonic skepticism and a working disagreement or a disagreement in fact, is a distinction between ontological skepticism and epistemic skepticism. Disagreement in fact about an alternative observation is an epistemic issue although someone who proposes an alternative observation is likely to draw ontological conclusions from his observation. But the epistemic skeptic does not challenge directly any assertion about what is in the world; rather, he challenges the validity of the observation. Indirectly he may challenge the validity of some conclusion about ontology drawn from the observation (Some of Hume’s arguments fall into this category).

There is also a difference between historical Pyrrhonic skepticism, which is logical (“There is no truth”) and Cartesian skepticism, which is epistemic (“There is no certain knowledge”). What is at issue here is Cartesian skepticism since the issue is the observational justification of a theory or assertion - an epistemic issue. But we are not, as I just noted, even concerned with thoroughgoing Cartesian skepticism, but only in cases involving non-standard perceptions. (A thoroughgoing Cartesian skeptic would ask how we know a given perception is non-standard.) The idea of radical translation, which I will elaborate on in a bit, allows us to draw a distinction between thoroughgoing Cartesian skepticism, whose advocate could be accused of speaking an untranslatable language, and alternative observations. The person who reports an alternative observation would resist our inclination to conclude that he is using his words in a different sense than we do. The situation would also presuppose a high degree of shared meanings, i.e. a shared language. An observation can only be understood as alternative if meanings remain the same.

A final distinction I would like to make is between observations about primary occurrences vs. secondary qualities. I borrow and bastardize some of Locke’s terminology here simply because it is the most useful (Later I will introduce a related distinction between first order and second order perceptions; you should notice that many perceptions of primary occurrences of qualities such as color are in fact second order perceptions). I shall use it informally without too much effort at rigorous definition.  A primary occurrence observation is, “The sample contains 5 moles of carbon.” When a primary occurrence observation is alternative, the basic sensations of the witness to that observation differ from the basic sensations of other people, perhaps most other people. An example is the observation that there is an angel in the room or that Jesus is hiding in the toilet. A secondary quality observation is, “He looks angry” or “This house looks like it was designed by Frank Lloyd Wright.” A candidate for an alternative second order observation would be one made by an individual when he looks at Lake Tahoe and he sees the same Lake Tahoe as other people (The basic sensations are the same), but to him Lake Tahoe shows signs of having been created by the Army Corps of Engineers. The eastern shore looks too straight. Saw grass is growing next to the water, something that never happens in nature. (For those for whom everything looks designed, this sort of evidence is unavailable. Since both a straight shore and a crooked shore would be signs of design, there is nothing that qualifies as looking, and therefore being undesigned. The observation can’t be falsified and for that reason it can’t be verified.)

There is some controversy that attaches to any primary/secondary distinction when it comes to object qualities (or to perceptions). For many the very idea of a primary quality is tied to philosophical essentialism, which, at the very least, needs vigorous defense. For others, all qualities (for reasons that I won’t go into here) are “secondary.” There is, so they say, no such thing as a primary or touchstone perception by which an object may be identified.  It is because of these controversies that I choose to treat my primary/secondary distinctions in this instance as merely a handy tool, a way of framing why we consider some observations as more trustworthy or more agreeable, so to speak, than others. The fact that there are observations that enjoy a high degree of “agreeability” is a primary observation if there ever was one. Nevertheless, my use of a primary/secondary distinction in this instance should not be counted upon as endorsing or assuming (or rejecting for that matter) any philosophical theory other than what is on the surface. Least of all essentialism in any of its forms. Once the distinction has served its purpose we can toss it in the closet till next winter.

What I am concerned with here are alternative observations, the sort that I have loosely defined as not shared by a significant number, perhaps the majority of other people. Obviously this definition is inadequate as it stands since a great number of people, perhaps the vast majority may, for example, perceive the world as designed. But the inadequacy of my definition is not critical. For one thing the principal argument against the perception of design was that its assertion leads to contradiction, a consideration that outweighs the actual number of people who actually think they have the perception. In a similar way most of Hume’s arguments are directed not so much against the perception of design as against the supporting lemmas used to draw the strong conclusion that some god somewhere designed the universe. Furthermore, I think it is worthwhile to isolate the situation where one person makes an observation that at least one other person does not share and then let them have it out mano a mano. The issue is to devise criteria for distinguishing potentially valid and invalid alternative perceptions, between theory shaking evidence and crackpot claims. One such criterion that has begun to emerge is significant disagreement, either disagreement with the beliefs of a large number of other people or incompatibility with an established corpus of scientific fact. Disagreement by itself is not definitive for otherwise we could never protect the rights of epistemic minorities or experience a scientific revolution. Rather, disagreement points toward a need for further investigation and the use of other, perhaps subtler, tools. The tools I sketch here will be for the purpose of assessing the validity of any alternative perception regardless of the actual number of proponents and opponents. In practice alternative perceptions meet with active resistance if not from the public at large, then from some significant segment of the population such as the scientific community.

I need hardly point out that the issue is not perception but observation. This is the difference informally speaking between the two: Observations are in a way more composite, more complex than perceptions. When we report on an observation we are comfortable with making some ontological assumptions and general assumptions about kinds, theories and laws. Observations mostly conform to the "normal" way we speak although that "normal" way may also be in the language of a particular science. We like to think of perceptions as the simpler parts of observations. We sometimes behave as if our observations are just groups of structured perceptions. But, in the opinion of some, perceptions are just a type of observation, and they may be right. At the very least it may not be possible to make perception reports that are not disguised observation reports. Observation sentences tend to be declarative descriptions of facts such as "Jenna and Nikki are doing a two girl scene." A perception report, if there really is such a thing would run something like, "I see a pinkish patch on top of a light brown patch." Perception reports are out of the ordinary artificial constructs. Oftentimes they are the product of specialized terminology created by philosophers. But that needn't always be the case. For example, if two spies are looking through their binoculars and they aren't quite sure what they see at a great distance, one might say to the other, "I see a pinkish patch on top of a light brown patch." The point of devising perception reports is to minimize disagreement as much as possible, to find a common ground from which hopefully more "normal" observation reports can be agreed upon. One spy may think he sees two girls on a beach while the other thinks he sees a walrus on wet sand. But at least they might agree that they see a pink patch on top of a brown patch. Some philosophers use this property of perception reports to create a basis for more articulated but less obvious scientific observations and theories.

When disagreement arises the issue is usually one of alternative observations and not alternative perceptions. We may not all see an obviously designed natural formation, but at least we can agree that we see the Grand Canyon or at the very least a bunch of brown patches. Some alternative observations, however, present themselves as perceptual anomalies. Either you see Jesus in the toilet or you don't. If you don't see Jesus in the toilet, you probably don't see a hairy brown patch that could be interpreted as Jesus in the toilet either. But our question mostly is: What constitutes a proper observation? Over and above that, what constitutes an observation that can be considered as valid for everyone? Finally on what observations can further conclusions be drawn and what is the relation between observations and the conclusions that are drawn from them? Obviously perception is an important part of observation, but the issues discussed here can be addressed without a prior conclusion as to the nature of perception.

Most issues regarding alternative observations arise from conclusions that can or cannot legitimately be drawn from perceptions or from acts of seeing-as. They are deductions from perceptions of (mostly) secondary qualities: The Grand Canyon looks designed, therefore god designed it. Some concern the perceptions of the secondary quality itself: The Grand Canyon looks designed. Can anything be said about supposedly primary perceptions? Or are they just “the given” We can approach this from a historical perspective. Alternative first order observations are justifiably rare unless they are the product of the introduction of an observational aid like the telescope or microscope. Alternative observations are most frequently what I have loosely called second-order, such as those that involve a second-order impression on an observer or occur in a matrix of meanings and facts that are already distinct from first order observations (It is important to view certain matrices as distinct and not necessarily just higher order. The art historical matrix of observations about sculpture is distinct from and not just added to observations such as, “There is a piece of marble in the room.”)

It would be unacceptable to treat primary observations as unassailable since that would destroy any basis for subscribing to or rejecting a belief. For one thing I could simply say I observe David Hume on a mountain top and he is obviously the king of the universe and no one has a right to doubt what I see. Luckily the design theory fiasco itself provides us some tools for singling out irrelevant or invalid observations. One tool is a kind of methodological consequentialism. If conclusions are or can be or must be drawn given an observation as an assumption, and if for one reason or another these conclusions are unacceptable, then if there is also a sound basis for doubting the merits of the observation, there may be sufficient grounds for rejecting the observation. Unacceptable consequences include logical contradiction or moral indefensibility in the conclusions that are or can be or must be drawn from the observation. Many alternative perceptions have no consequences for us at all and don’t necessarily call for a response. Some, when indirectly communicated through works of art and literature convey emotion or pleasure. It is when an the assertion of an alternative perception entails a logical fallacy in some significant way (outside the realm of fantasy) or implies that we must be forced to do something or compelled not to do something for which there are no other grounds that it should be opposed, questioned and ridiculed.

Interestingly enough Nelson Goodman (pp. 99-100) suggests precisely the idea of consequences in addressing a kindred phenomenon such as is exemplified by the assertion that two colors observed at different times are the same color. Since past presentations, as he calls them, cannot be revived for the purposes of comparison, such assertions are unverifiable. He calls these assertions "decrees" (I like the ecclesiastical ring of that term. It ties nicely with the theme of this essay). Goodman's suggestion for testing decrees is to look for the consequences of their acceptance. If a decree results in inconsistency, for example, we would feel justified in abandoning it.

The Consequences Test joins the Agnosiac Fallacy and the Universalizing Fallacy as tools that can be used to defeat arguments from the observationally given when conclusions are drawn beyond the mere assertion, “I observe this, here, now.” All three provide a way of looking under the hood of arguments from the given to see if conclusions are drawn that go beyond the given.

Two philosophical theories have tried to carve out a realm of observation whose contents are somehow essentially immune from non-Pyrrhonic challenge. None of the observations belonging to this realm can be alternative since, by definition, there is nothing that they can be alternative to. They are the touchstone against which all other observations are evaluated for validity. Their immunity consists partly in the fact that these observations cannot be shared at least by a majority of other people, or in the end any other people. Only one person can or cannot make this kind of observation. Only one vote counts in the kingdom of the private sensation. Obviously these theories are sense data theory and the inutionist-introspective element of Husserl’s phenomenology. I will concentrate at this point on the former, although what I say applies pretty equally to the latter also. Sense data theory (or mutatis mutandis the theory of protocal statements), most commonly associated with some versions of logical positivism, is one empiricist sort of philosophical theory of scientific observation (and not as we may sometimes think, coextensive with empiricism). According to sense data theory, our “immediate” sensations have a special status that confers upon them a kind of incontrovertible certainty. Equally the data of our immediate sensations are undeniable and reports about the data of those sensations are about as true as any report can be. Because of this certainty, our immediate sensations are suitable to serve as the foundation for scientific theories or models. According to the well known example, I may not be certain that I see a bent stick. In fact my belief that I see a bent stick may be in error. But I can be certain that “I” (an indexical that must be further indexed by time and place) have a sensation of the sense datum “bent stick.” Some, who justifiably feel uncomfortable with references to physical objects, whose ontological status may be subject to doubt, prefer to substitute terms like “patch of green” for “bent stick.”

Since its first triumphant appearance on a train between Cambridge and London somewhere sense data theory has had something of a rough time of it. In a nutshell, the theory was betrayed by the language used to describe a sense datum (and of course without an observation language no science). One of the errors of sense data theory was the assumption that sense-data and perceptions occur in a significatory vacuum, that for each mental intuition of a sense datum there was a specific sense datum that could be grasped (and perhaps named or understood) without indirect reference to the entirety of or some significant portion of what the perceiving subject knows or has otherwise experienced. Thus, even granting provisionally the certainty of sense data sensations (to use a physiological term) or perceptions (to use a psychological term - the two are not always clearly distinguished by sense-data theorists), there is no clear singular one-to-one correspondence between a sense datum (something independent of the observer but not necessarily physical), the perception of that sense datum (something the observer does) and the observer’s verbal report of the sense-datum relating sensation. By the report stage the observer’s language and category system have intervened. (Interestingly structuralist phonetics provides a model for understanding this error in its insistence that sounds and phonemes do not perform their significatory function by simple singular reference to something signified. Rather the significance must be elicited from the place the sound or phoneme occupies in a more complete system of signification. The sense-datum/perception relation is a great deal similar.)

Another interesting point is that Russell’s early version of sense data theory – a version of which logical positivism could be called the comic book version – was an attempt to express an important insight. Part and parcel of Russell’s theory was that the only things that could legitimately be given proper names (i.e. names that have a singular one-to-one correspondence with the object they name) are our sensations (or, ambiguously, the data that are the content of those sensations). Everything else is a construct more correctly expressed in language by means of definite descriptions. A little reflection will show that Russell’s idea actually comes down to saying much the same thing as the language-based criticisms of sense data theory. (Popper’s falsifying observations are not an essential part of a foundationalist theory of knowledge and so they don’t fall within the scope of this discussion. Clearly, they are alternative in one sense: They conflict with observational predictions on the part of a canonical theory. But equally clearly they are not alternative in another sense: A genuine falsifying observation is shared.)

Since both sense data theorists and phenomenologists (at least those who didn’t deviate too far from what Husserl actually said) regard themselves as the legitimate heirs of the philosophical approach that has with a degree of vagueness been call “empiricist,” that approach finds itself troubled in the muddy waters of observational unreliability. We quite often don't "see" the same things (aside from the fact that none of us supposedly have the same sensations), that seeing as a mental phenomenon is conditional in the sense it probably is never pure, that it is always influenced by some factor or other. One might go so far as to say that perceptual ambiguity strikes at the very heart of empiricism, and by extension, at the scientific method itself. The proper response, of course is not to become defensive about one’s observations, but rather to investigate observational discrepancies dispassionately. So quite a bit is at stake. At some point it will behoove us to let the scientific warriors do real battle and leave the creationist adolescents to play with their watches. The only point that should be retained in the context of this essay is that, because of essential ambiguities in sense-data reports, there may very well not be any observations that are immune from Cartesian skepticism. And ultimately epistemological foundationalism provides no help in distinguishing legitimate from bogus alternative observations.

Put in a different way, the foundationalists are correct in asserting that certain propositions serve as the starting point from which other propositions are derived. They are wrong in believing that some mysterious property called "certainty" or "necessity" could adhere to those basic or foundational propositions - if only because the basic propositions may be overturned in time. But instead of rushing to an immediate and as far as we know so far self-contradictory relativism, it may be preferable to try to work out criteria for assessing the at least provisional validity of those propositions that could constitute a starting point. There may not be certainty, but there is faute de mieux.

Moreover, the issue I am addressing here is not the Hegel-Schlick-Quine (Phew!) issue of whether any observations such as the observations that are the content of sense-data reports are truly immediate or truly directly referential, i.e. impregnably valid. Rather I am concerned with more of a working situation wherein we acknowledge that some observations are as valid as they can be for all we know, and, that being the case, what do we do with anomalous observations, viz. observations that somehow don’t fit in or aren’t shared by other observers. What interests me is not the logical criticism of sense data theory or of the certainty of our individual sensations, but rather, how in fact a philosophical theory is supposed to handle alternative sensation reports (which are not immune from challenge although the sensations that they report are, according to sense-data theory, so immune. For, assuming the validity of the Law of Non-Contradiction, if your report (or belief) of what you observe entails a contradiction, then what you report (or believe) cannot be what you actually observed.). For the most part this doesn’t cause problems because agreements about what sensations we have and share are or appear to be overwhelming (at least within a culture; many think that external cultures are more or less satisfactorily handled by arguments for radical translation).

But the claim by goddists just to “see” design in the physical universe unearths the working problem in all its worm eaten finery. What if someone reports that he just sees an angel in the room that you don’t see? What if exactly half the respondents in a homogeneous group of psychology test subjects report that they see a red patch and the other half report seeing a blue patch? What if the observers of the experiment also cannot agree about the color of the projected patch even though everybody agrees about the difference between red and blue outside the confines of the experiment? Radical interpretation is not a satisfactory solution to this problem precisely because in the working situation it is assumed that all the observers share the same meanings, i.e. it is assumed that everybody means the same thing by “angel” or by “patch” or “red” or “blue.” To argue that the design theorist just means something different by “designed” is hardly a convincing response. We must assume that his meaning matches our own sufficiently so that, where it does deviate, we can prove him wrong. This is what I did in the previous section. We share a meaning with Paley when we both agree that a watch is designed. It is only because we do share this meaning that we can go on to argue that his supposed proof of the designedness of the entire universe eliminates the possibility of distinguishing designed from undersigned things within the universe and thus eliminates the intuitive basis of his proof. (This is the nightmare scenario for Davidson’s oft disavowed but relentlessly recurring intimation that only meanings change, the facts remain the same. After all those Renaissance cardinals could have easily disposed of the mythical Galileo by claiming radical interpretation, viz. the meaning of “round” in his object language is different from the meaning of “round” in their metalangauge.)

Things like that never happen outside a Philip K. Dick novel you say? Well, the science fiction aspect of the latter examples is tempered by their kinship with real world conundrums such as may be posed by design theory. We can and should dismiss design theorists as kooks (They are kooks). But alternative perceptions deserve more analysis and not dismissal (Clumsily applied, the fact/meaning distinction is a dismissal). So do what may be called innovative perceptions. The latter as we saw constitute the woof and warp of the human sciences and literary criticism. “First order” perceptions are sitting ducks for a Davidson type analysis: If you say snow isn’t white, you probably mean something different by “snow” or by “white” than I do. “Higher order” perceptions (Let’s use the concept of perceptual orders in an informal way here and leave attempts at a rigorous definition for a different time and place) such as “The Master of Flémalle’s pictorial space is planar and not container space,” may not be shared upon first report, even though they do depend on shared meanings as much as any debate about the color of snow. It may take persuasive description on the part of the art historian to influence  others to share his observation.

In The Brown Book (pp. 138 ff) Wittgenstein describes an almost classic situation of alternative observation. Someone is "taught" the meaning of the words "lighter" and "darker" but immediately proceeds to apply these terms to the vowels of the English language in such a way as to lead the writer to believe that the instructee actually means something different by “lighter” and “darker” than what he had supposedly been taught. “We may,” says Wittgenstein, “be inclined to treat this case as some kind of abnormality....” (p. 139) He approaches the case in terms of what Davidson would call radical interpretation, but with a difference. Instead of stressing the interplay of meaning and fact Wittgenstein tries to show that apparently anomalous usages are no more than what we describe, a sequence of actions that would not correspond to the way we might act, but are not for that reason the product of a hidden process. One might say that there is no need for radical interpretation at all; just go with the flow.

The issue Wittgenstein was addressing is not quite the same as the one that concerns us here. Wittgenstein wanted to show how it is possible to describe and so understand differences of meaning without appealing to anything beyond what people say and do. “Are you distinguishing between the sense in which he used the word and his usage of the word? That is, do you wish to say that if someone uses the word as B (the instructee - WD) does, some other difference, say in his mind, must go along with the difference in usage? “ (Ibid.) Wittgenstein's point is that the "abnormal" situation where B applies “lighter” and “darker” to vowels without coaching need not be understood as a difference in meaning at all. This behavior is simply part of his meaning.

Obviously what concerns us here is the fact that there can be abnormalities at all, that we are faced with alternative observations. And, in choosing an obviously secondary quality like shades of lightness and darkness, Wittgenstein paints a clear picture of such a perplexing situation. I just want to add that Wittgenstein's conclusion - or one consequence that may be drawn from it, viz. a sort of observational and significatory ecumenism - may very well not be serviceable in many if not most situations. Let us grant the basic conclusion, which I think may very well be true, that we don't need to appeal to mental states to describe anomalous differences in what people mean. Nevertheless our need to account for hallucination and observational error can live comfortably with Wittgenstein’s anti-mentalism. If crazy Mary says she sees a gremlin in the fireplace when we don't, it is not always fruitful to describe the situation as a case of difference of meaning whether that difference be understood in a mentalist or an anti-mentalist sense. Mary might just be hallucinating. Likewise it is most often better to treat the assertion that the earth is flat not as a difference of meaning but as a simple observational error. And there are cases where we might feel a great deal of inclination to treat anomaly as error. If B were to say “The letter E is lighter than the letter U. Therefore you are my slave,” we may be tempted to demur.

So, are there grounds for dismissing an alternative perception other than simply saying “I don’t see it” or presuming that the alternative observer might be using language differently than I do? And can the delicate embryo of an innovative perception be nurtured so that it has a fair chance of developing into something valuable and not die in the womb?

J.S. Mill was one of the first truly modern theoreticians of science and we can expect that he had a lot to say about observation and the justification of observation reports. Some of us think that Mill’s comments are far more relevant to the way science is viewed today than the detours represented by sense data theory and phenomenology.

One class of inductive fallacies that Mill identifies are what he (p.490) calls fallacies of mal-observation. These, of course, do occur or else anything that anyone perceives and claims to perceive (such as my vision of Jesus X butt-fucking Sister Agatha) actually obtains. Once Mill identifies the phenomenon of mal-observation, the problem becomes one of distinguishing proper observation from mal-observation. I hope I refined Mill's idea a little bit by introducing the notion of alternative observations. We also need to understand that the proper complement of the concept of alternative observations is that of common observations or commonly held observations. This distinction is different from that between valid and invalid (proper or mal, to use Millian terminology) observations. We need alternative observations. They keep science moving just as genetic mutations drive evolution. And some disciplines like the humanities live almost entirely on alternative observations, which might in this case be termed original insights or perspicacious readings. But there also has to be a mechanism for distinguishing valid from invalid observations. Sadly not every nut can have his day. Remark that, just as alternative does not equal invalid, commonly held does not amount to valid. Improperly understood, the observation that the sun rises in the east is commonly held but invalid. This latter distinction is important for those who want to combat things like prevailing moral codes.

Mill (pp. 408 ff.) provides a rationale for distrusting a perception that violates an established law of causation. The basis for or sole evidence of many non-standard or alternative perceptions is the testimony of the person who claims to have (had) that perception, not any experience of our own. Our only experience is the hearing of his report. Obviously we rely on witnesses a lot. They are pretty much all he have when it comes to (currently) inaccessible events such as past events. But, as we know from The Crucible, some witnesses are better than others. This produces another test for evaluating alternative observation reports. Could the witness be motivated to report something less than the absolute truth? On pp. 506 ff. Mill draws up a list of causes or motives of simple mal-observation, a marvelous contribution on his part to motivational psychology. Mill's point applies not just to inaccessible first order perceptions but also to observation reports of unshared second order perceptions. The person who reports "just seeing" a secondary quality (such as designedness) should be treated as we would any other witness. What Mill has to say is clearly a theoretical refinement of Hume's little dustup with the disciples and other "witnesses" to miraculous goings on. (Cf. Mill pp. 513 and 514 on marvelous tales and visions of god). Hume by the way was not alone. Enlightenment Bible scholarship particularly in Germany picked apart the so-called New Testament just as thoroughly as Spinoza, Bolingbrooke et al. demolished the Old Testament. Cf. D.F. Strauss for a summary of the research.) What Hume and Mill said about miracles applies to second order observation reports where the validity of the primary observation is not at issue. It is not quite the same as what I call the Consequentialist Test for alternative observations, namely the test for conflict with shared observation or logical or behavioral unacceptability. Let’s give it its own name: The Consider Your Source Principle. In the case of seeing-as (the pope in the pizza) the claim is to be evaluated by balancing probabilities. Is it more probable that the perceived gestalt is the work an external agent or that Maria had been conditioned by years of indoctrination to look for such things and that all those bishops and prebyters and shit  were motivated by the chance for a power grab, an the opportunity to score a goal for the team? (All those freshly scrubbed young seminarians are encouraged to trot out Bayes at this point. Let's face it. The guy can't even stop spam. And we're supposed to believe that he provides a "mathematical" proof that we should believe some bearded wonder's claim to have seen miracles performed? Please.)

By way of aside I hardly need recall Lucretius’ perspicacious comment about sensus communis:

…cui nisi prima fides fundata valebit, haud erit occultis de rebus quo referentes confirmare animi quicquam ratione queamus. (I. 423-425)

Our issue, of course, is to find a way to distinguish genuinely shared sensations or perceptions from those we call alternative, and, within alternative perceptions, to distinguish between those that are or may be fruitful from the downright untenable.

The interconnectedness of observations provides one clue. Observations (or, more cautiously stated, observation reports) are not isolated phenomena. They do not occur in one-to-one correspondence with some fact about the world without a network of links to other observations. Every time an observation is reported, both parties to the communication presuppose the truth of an unlimited number of other stated observations. And the new observation may conflict with or not sit comfortably with or not jive with the others. The traditional scientific problem of anomalous and falsifying data is an instance of this. To what point is the anomalous data ignored and at what point does it require a revision of the prevailing model?

The interconnectedness of observations cannot simply be thought away. A thesaurus of shared observations is a necessary precondition for the existence of a community of "rational entities," to steal a term from Kant. It is just as necessary as the shared recognition of facts and meanings. The necessity of these preconditions is particularly clear for the scientific community which is constituted in a large part by acceptance of a number of observation statements and theories (or models). Yet the thesaurus of shared observations is not and cannot be a locked box. Periodically observations are proposed that I have characterized as alternative, that is they are novel, previously unstated and, in the most interesting cases, incompatible with theories accepted by the rational community or with other observations accepted by the rational community or with both. Alternative observations cannot be dismissed. They have proven to be the motor of progress in the physical sciences and quite often the culprit behind what we regard as scientific revolutions. Equally the humane sciences or humanities are powered in large part by alternative observations in place of experiment or induction. To characterize Roman art around the time of Constantine as the product of a different Kunstwollen or artistic intention and not simply as the result of degraded technical skills is to have proposed an alternative observation and not an inductive conclusion. The proposal was presumably based on lengthy acquaintance with the objects in question the result of which was a kind of apophansis, a seeing of the world or a portion of the world in a different way. To make the observation pass from the status of alternative to that of shared and accepted, Riegl used the technique of persuasive description, for no amount of induction or prediction could establish his insight as valid. However, not all alternative observations are worthy of acceptance. The claim on the part of Paley and other creationists to the effect that the entire universe “looks” intentionally designed is an alternative observation so completely without merit that it falls into the crackpot category. So it is a real philosophical issue not restricted to the theory of science as to how to distinguish valid from invalid alternative observations. I will try to suggest a few techniques here and to point out fallacies whose commission on the part of the proponents of an alternative observation would render that observation invalid.

Obviously you cannot simply say, “I see an angel in the room and you don’t and that’s that” or “I’ve got my magisterium and you’ve got yours” or “I can just plainly see that god made the Grand Canyon and you have no right to challenge what I see.” These observations occur within a thick nexus of shared meanings, other observations mutually held as veridical, behaviors and behavioral dispositions, and, tellingly, moral injunctions and allowances, that mean that they are not innocent differences of opinion (or rather differences of vision). If an alternative observation conflicts with accepted observations, its truth may involve the unacceptable rejection of those observations or indeed a radical revision of what is considered logical consistency. Or else its acceptance may ultimately result in the imposition of moral injunctions or calls to action that we find repulsive. It is this way in which an observation gets along or quarrels with its neighbors that allows us to address apparently irreducible perceptual conflict, to deal with the apparent conversation ender, “I just see things differently.” The resemblance between those who claim an alternative set of observations (including those who appeal to “revelation”) and the Pyrhhonist is not fortuitous. In either case there appears to be no shared ground whatsoever, no way of breaching the unshakeable fortress that lies beyond the reach of rational argument, logical consistency or a corpus of fundamental truths about what we perceive. In either case, that appearance is not really true or else one party could not even be understood by the other party. It becomes like the background noise of an urban landscape, incapable of even making its skeptical or dogmatic point. But as soon as anything is admitted as shared, the potential for a clash of observations or unacceptable consequences explodes like a cherry bomb.

There is another technique for evaluating an alternative observation which belongs more to what we might call philosophical working tools instead of solid unflinching argument. Let’s call it the Consider the Source Principle. It consists in bringing the character or background of the witness or simply the relative reliability of the report into the equation. Obviously witnesses with a prior record of mendacity or discernable self-interest (such as priests or goddist ministers) will not make stellar sources for trustworthy observations. But the reliability of a source can be called into question without much reference to the character or the background of the witness. We tend not to believe in reports of alien abductions even when we know little or nothing about the person who made the report. (Such a report by Richard Feynman, of course, would give one pause.) Hume’s argument against miracles to the effect that the likelihood that the witnesses to the miracles were lying, outweighs the likelihood that a real miracle really occurred is the classic example from the history of modern philosophy.

This principle is clearly dangerously susceptible to rhetorical misuse. Cicero provides a prime example of its application and misapplication in one of his speeches in defense of a political figured accused of extortion. He goes so far as to shape his tactic into something of a formal theory of dialectic, suggesting that, where the opportunity for real argument is lacking, one could attack the character of the witnesses: In hominem dicendum est igitur, cum oratio argumentationum non habet. (Pro Flacco, p. 466) What Cicero did in fact was launch a vicious racist diatribe slurring the Greek nation as only someone with his mastery of Latin rhetoric could really do. The Greeks are unreliable witnesses for the prosecution, he asserts, because they are lazy, shiftless, lying, dishonest, money grubbing, gluttonous purveyors of disgusting sexual perversions. The only possible recourse is to disqualify every bit of their testimony. The accusations are, of course, ridiculous on their face and could only hope to have an effect on a Roman audience already prone to hold such beliefs about the inhabitants of their subject provinces. The rather lame attempt to distinguish the upright Athenians from the Asiatic Greeks ignores the fact that Greek science was born in Ionia and not Athens. The irony that the British would many centuries later come to hold precisely the same opinion about the impoverished Italians could actually serve to illustrate another principle that could be called The Top Dog Gets to Write the History Principle. (In Cicero’s defense he did consider in hominem argument as something of a last resort. He did not advocate applying it in cases where you have actual good argument for your side; he meant cases where there were no grounds for a rational argument one way or the other for assessing the facts beyond the bald assertion of a single witness or closely associated group of witnesses.)  

The example of Pro Flacco shows just how volatile and susceptible to misuse the Consider Your Source Principle can be. But it can be of use particularly where there is a legitimate conceptual framework for defining disqualifying character traits or states of mind. And where the ascription of these traits can be accurately and dispassionately applied to a witness without prejudice or emotional bias. One such framework is what Nietzsche called the cui bono test. If a witness stands to gain from acceptance of his or her testimony, then there may be grounds to disqualify it. But the gain must be real and evident to the witness even if it may not be immediate or even completely tangible. A witness can’t be dismissed just because “everybody knows” he’s a lazy shiftless bum. Another potentially legitimate approach is to consider whether the witness has some sort of axe to grind, whether he or she is seeking vengeance or is acting in the service of some political or social program. The first century CE purveyors of the Jesus myth most likely fall into the axe grinding category, while modern day preachers are suspect because of the tangible (usually financial) rewards they may reap even if those rewards are no more than heightened self-esteem. An actual history of misrepresentation or false report is another potentially legitimate framework for dismissal. When applied to the issue of reliable observation, the Consider Your Source Principle does suggest the valid question of whether someone who proposes an alternative observation could on other grounds be classed as a liar or a fantasist.

A similar test for the validity of observations is suggested by psychological research on the effect of motivation on perception. (Experiments in Visual Perception, pp. 365 ff. esp. pp. 389 ff.)  We might call it the Principle of Dispositional Distortion. The criterion is that if there is an existing body of evidence that certain dispositions can alter perceptions and if someone is disposed in that way, then his perception and consequently his observation could be altered. Obviously this criterion like the others is a working tool. A thoroughgoing skeptic could argue that the existing body of evidence is itself subject to doubt as to its accuracy as an observation or set of observations. But, since this criterion is a working tool, we can safely make the assumption that the existing body of evidence is part of a commonly held set of beliefs or part of the scientific canon. If the person whose observation is potentially motivated rejects the evidence from psychology, he is making a different or larger point. He is not simply proposing an anomalous observation as valid, he is also either raising a philosophical issue or rejecting all or part of the evidence surrounding the effect of motivation on perception.

The Consider Your Source criterion and the Principle of Consequentiality are two complementary sides of a single methodological coin. One principle enjoins us to look closely at the origin of an alternative observation; the other suggests that we assess what would follow if it were true, not only other “truths” but also modifications in our behavior. Both are useful means of evaluating whether or not we should reject an alternative observation when there is no recourse to logical refutation or falsifying observations.

Let us summarize some of these debunking tools which in my opinion deserve at least as much respect as Mill’s principles of induction.

The Principle of Consequentiality. Does acceptance of the observation conflict with a law of logic? Does it lead to other consequences (including moral consequence) that we find unacceptable?

The Principle of Overwhelming Acceptance. Is a conflicting observation or set of truths nearly universally accepted?  This principle does no more than restate that a given observation is alternative. Conflict with overwhelmingly accepted truths does not in itself constitute a reason to reject an alternative observation. Indeed the Principle of Overwhelming Acceptance should be applied to evaluative statements only with a great deal of care. If it applied to evaluative statements, then rational dissent would be excluded. Socrates would always be wrong by definition.

The Agnosiac Fallacy. Does the observation obliterate the distinction between what is observed and what is not observed?

The Universalizing Fallacy. Do conclusions drawn from the observation attribute some property to the complement of the set of all the things otherwise considered to have that property?

The Consider the Source Principle. This is one of Hume’s signature contributions to the scientific method.

Similar to the Hume Principle is the Principle of Dispositional Distortion.

These mostly negative tests for judging the validity of observations constitute something of a mini canon of tools for evaluating the validity of primary observations especially in cases where those primary observations conflict with existing theory or other observations. Let me recall that primary observations and the propositions that report those observations (propositions that some have called “protocol propositions”) are in a sense underived, if anything can indeed be truly pure and not derived from something else. The best criterion for a primary observation is, I suppose, whether it can be contravened by some form of traditional reasoning largely divorced from appeal to other observations. If such a contravention is difficult or nearly impossible, then the observation has a claim to primacy; it serves as a basis for other beliefs.

Somewhat grandiosely I like to compare my little canon to Mill’s canon of techniques of inductive reasoning or even to the canon of deductive logic. It addresses a much overlooked and seemingly untouchable epistemic thesaurus, viz. our initial contact with the world. These tools are in my opinion almost as important as Mill's principles for valid inductive reasoning. They are more important in some ways, for canons of valid inductive and deductive reasoning assume some starting point of valid assertion, a starting point that is usually - one is tempted to say always - observation based. The criteria for observation evaluation are meant to sort out what can even so much as serve as a starting point for valid reasoning.

IV. Epilogue

Of all the so-called proofs for the existence of Godot only the creationist swindle seems to persist despite the best efforts of modern science, much the same way as the common cold continues to annoy us even though smallpox has been largely eradicated. These days the ontological argument is given barely a passing nod and lives on only in the troubled imaginations of charlatans like Alvin Plantinga. Not even papists make much mention of the Five Ways first bruited about by the Dumb Ox. Perhaps the creationist cult survives because it stimulates the dulled brains of the illiterate with its perfervid imagery of rose red sunsets and happy butterflies. “If god didn’t make that there Grand Canyon,” cries the toothless goober, “then I am a monkeez onkle.” Unfortunately the American hinterlands are overrun with just such simians – a curious evolutionary inversion that merits serious biological investigation. Where philosophy and science haven’t quite done the trick perhaps a dose of fuckosophy – the legendary crystal meth in the intellectual pharmacy – will help obliterate the nuisance.

By the way, whatever happened to Maria and her pizza? Minute inspection by interfaith theologians and Vatican officials of the pizza’s intricate sausage pattern revealed a disturbing image that the media felt was best discreetly disregarded: Pope Armando was butt fucking her son!