Truth

I. Introduction 118

1. Some background reflections. 120

2. Some general ideas aimed at in the explorations to follow. 123

3. Now we have to dislodge some metaphors. 128

(a) The first: The matterless mirror. 128

(b) The second: The structure of the mind. 129

(c) The third: The invisible subject. 130

II. How Can Things Be Parts of Thoughts? 136

1. We do not live in representations. 137

2. Some remarks on representationalism. 139

3. Internal to external awareness. No. 143

4. Aristotle and Descartes: Nothing physical becomes mental. 145

III. Two Defeating Models 155

1. First defeating model. 155

2. Second defeating model: Attributional-intentionality. 159

3. Now what can we conclude from all this? 161

IV. Bad Global Theories of Truth 162

1. No correspondence. 163

2. Coherence. 171

Chapter 4: Outline Page 2

IV. Bad Global Theories of Truth (cont.)

3. Pragmatism. Historical vs. empirical pregnancy. 173

4. A transition. 175

5. Somethimes the truth is what counts as so. 176

6. What about "eliminative," "redundancy," and "disquotational" analyses of truth? 177

7. Local roles. 178

8. Emotivism and other add-on theories of truth. 180

9. Parasitic locality. 180

V. What Does the Failure of All Those Theories of Truth Tell Us? 181

Chapter 4

Truth

"...proud man,

Most ignorant of what he's most assured,

His glassy essence."

I. Introduction

The trouble about "truth" comes from mistakes about thinking, especially from construing features of sentences to be features of thought, and vice versa, and from expecting a single basic analysis of all propositional uses of "it is true that" and "it is false that." There is a family of general notions of truth and falsity, but the expressions "is true" and "is false" are so entrenched in diverse particular practices, like arithmetic, logic, sciences, legal processes, and even sub-processes like various forms of contrary-to-fact speculation, that (i) there can be no single correct analysis to cover all the cases; and (ii) there is no "core" notion with only "accidental" differential features [see C. Wright, 1992] because the meanings in different practices are genuinely different and sometimes have different "overflow" conditions. For instance, sometimes bifurcation is required, sometimes a commitment to no gaps, and sometimes one or another form of cognitive accessibility. (iii) In a variety of discourses, cognitive accessibility requirements range from "warranted assertability" to "constructive provability" "pragmatic verifiability," varying in whether such a requirement is part of the linguistic meaning of "is true" or part of the "overflow" signification. (iv) The standard theories of truth, (e.g., correspondence, coherence, pragmatic, redundancy, disquotational, etc.) either fail to be generalizable into a global theory, or fail to explain what they purport to. For instance, the so-called "correspondence" theories, that are "same thing" theories, fail to explain how there is a truthmaking identification of what is said with what is so. On that point I intend to supply the missing explanatory feature [see Chapter 5].

Thus this chapter is designed to require an account of how real things and real situations become constituents of our judgments (provided in the next chapter).

The notions of "the false" diverge enough from the notions of "the true" and have been so little and so unsuccessfully addressed in prior theories that I need a separate chapter for that [Chapter 8].

The overall outcome is that "true" is analogous in meaning, subject to semantic contagion, is not always eliminable even as a word, much less as a relation, and is a genuine predicate (without any platonic abstracta). Sometimes the notion involves cognitive elements; sometimes, it applies with gaps; sometimes with "not true" not being equivalent to "false." Most importantly, when we take the problem of explaining truth and falsity to concern intelligent human judgment transforming animal awareness, the locus of what is to be explained shifts, with issues about how real things belong to thoughts becoming central, and issues concerning whether "is true" is eliminable by disquotation, and issues about the liar paradoxes becoming secondary and resolved as consequences. Moreover, we are forced to reconceive physical reality and animal perception so as to account for the presence of real things and real conditions of things in animal and human awareness.

1. Some background reflections.

The 17th century adjustments to accommodate the deterministic, mechanical and corpuscular reconception of the physical world mutated and endured to become background certainties even now: representationalism, evidentialism, a mind-external-world gap, phenomenalism, the idea that sentences bear truth or falsity, even that "is true" and "is false" are redundant or not genuine predicates, that language is indispensable to understanding, and eventually, materialist idealism.

The notions of truth were affected when the rejection of intrinsic forms and multiple real natures in favor of a single mechanistic universal res extensa obviated abstraction as the basic and constant operation of human understanding and changed what was meant by "idea." For a mechanism that holds all matter is deterministic and governed by mathematizable mechanical principles (Descartes) and is all of one nature obviates the constant abstraction of physical forms into concepts. Secondly, rejecting a "dematerializing" understanding that operates on phantasms (sense appearances) to apprehend forms as held by Aristotle, Aquinas, and later scholastics, encouraged a retreat from regarding human understanding as a transformation of animal awareness toward Cartesian-Augustinian duality of a separate immaterial soul, all of whose ideas are innate (as to content), with an inner theater of experience, and gradually tended to treat "ideas" not as conceptions but as sensory contents, impressions, and images (e.g., Berkeley, then Hume). Further, the denial of animal awareness by Descartes, and its subsequently being treated peripherally, disregarded the continuity of cognition among animals and emphasized the truths and falsities of human judgment, with the result that animal perception (e.g., of sounds) became funged with intelligent perception (e.g., of tunes) and with the notions of truth and justification. Further, the notion of truth became associated with expressed truths, statements, and, then, with the (re)invention of sentential logic, became associated with the expressions themselves, sentences.

The "linguistic turn's" overemphasis on language encouraged convictions about truth and meaning that have now assumed the role of certainties. For example, most philosophers think "truth" is a semantic notion, a feature of sentences, or at best, of statements or propositions, and treat formal necessities, like "3 times 5 is fifteen" and "the bisector of the angle at the apex of an isosceles plane triangle perpendicularly bisects the base" -- which do not have any real cases -- as "the real thing" as far a necessities go, even more real because more precise than "necessities of nature," like "iron is magnetizable", "people can't digest stones" and "all integral mammals have seven cervical vertebrae, whether they are giraffes, humans or mice" -- which are paradoxically, denied to be necessary and considered only firm probabilities, or left without explanation or said to be conceptual or linguistic inclusions (as if the reason people can't digest stones is because of how we think or talk or as if the number of vertebrae in the neck were just a matter of how we count them).

Truth is an intrinsic feature of (some) thinking. It does not consist in a match-up or working-out or hanging-together of sentences or even of thoughts, but in the sameness of what we think and the being of things, as will be shown later. The content of true judgment is how things are. It will only mislead to think of the sameness as a kind of isomorphism, or a kind of congruence where what we think or say coincides with what is so; I am going to explain a content-identity account instead.

To get to that outcome, I (1) criticize prevailing metaphors and models of human perception and (2) show that no parallelism of thought and reality can allow an acceptable analysis of truth regardless of how we deal with the alleged "cognitive access" elements of truth and with the supposed requirement that there be an independent fact of the matter; and (3) give a positive account of abstraction by which real things become constituents of thought, just as real things are present perceptually to animals. The third step occurs partly in this and the next chapter and leads on to further discussion of the nature of intelligent animal thought [Chapter 6] and of the structures of the material world [Chapter 7].

Do we say there are no brain states that are representations? Not at all. There must be such states in all complex animals. But they, and functions among them, are not constituents of thoughts (or even of animal cognition), no matter how hard identity theorists try to make that come out. They are not even productive causes of thought or human perceptive action; and, one by one may not be necessarily connected with thought, perception or feeling, either. Rather, just as you cannot have a written message without writing and just as that requires a distribution of ink or other molecules on a page, you cannot explain the content of the message from the ink molecules on the page, but only the reverse; so the brain states are to be explained downward from the cognitive states, even though some changes, defects and deficiencies in brain states cause cognitive deficits, just as my erasing a word and putting new ink marks in its place may change the message.

Whether certain sorts of brain states are part of the "overflow" de re necessities of human thoughts is another matter, to be considered as we return to the status of brain states from time to time, though it should be obvious that the optical brain system is de re necessary for human intelligent seeing even though not conceptually contained in the notion.

2. Some general ideas aimed at in the explorations to follow:

(a) Though the platform of human cognition is animal awareness, (i) human cognition is a constant transformation of it by abstraction of what it makes present, and (ii) the principle of human awareness is not the same, but contains the abilities of animal awareness eminently.

(b) Nothing in nature knows what it is, but only humans are mistaken on the matter. Every answer the philosophers have offered is wrong except at the very trivial generality that we are a zoon logicon. That's because of man's glassy essence: we are animals, but more. But how much more? So the question of whether the principle of animality (the organization of carbon-based molecules by which we are able to sense, perceive, feel and act as unities) is really distinct from the principle of rationality (of understanding, willing and pursuing the goals of a rational creature) or is contained eminently by the latter (which is of an entirely different order from that of animals in general) is of paramount importance. Humans are capable of activities that cannot be performed physically [Chapter 6], and exercise animal powers from an immaterial psyche.

(c) Truth, globally, is right thinking, detectable as right to thought alone, as to what is or is not [cf. Anselm, De Veritate]. What I think is true when what I think is what is. "True" is a plastic notion, adopting distinct meaning elements in diverse discourses and requiring diverse analyses to articulate those diversities of meaning. Moreover, there are "overflow" differences in what is involved in truth or falsity that do not show up as differences of meaning, though they show up in differences of usage: thus sometimes every proposition is assumed to be true or false; other times, nothing is regarded as true that is not cognitively accessible (e.g., constructively provable mathematically), and there are truth-value gaps where verification is not in principle possible, or where there are no accepted rational standards for comparison: "The industrial revolution was more destructive than Colonialism." Sometimes there can be truths and falsities even when we do not acknowledge any independent fact of the matter and other times not. Thus, I think M. Dummett was right to say "meaning is entirely determined by use"; but that has to balanced off with "but use is not entirely reflected in linguistic meaning" because there are additional practices of reference and additional overflow conditions for applicability (signification) and there are ways expressions are used to modify actions that belong to the pragmatic traction between discourse and action rather than the linguistic meaning.

(d) Judgment flows as seamlessly in content as sensible awareness does. We scab it over with sentences to match it up with pairings from reality that we call facts. So generally, I prefer to speak of commitment rather than belief, because "belief" has become too readily associated with sentential expression, and, thus, with discrete packaging, whereas judgments can be simultaneous, overlapping and continuous, and mixed between reality commitments (existential judgment), characterizations (predications), and identifications. As will become clear, real things and their conditions are elements of both animal awareness and human intelligent awareness and constituents of our judgments. Yet only intelligent awareness involves truth or falsity.

(e) No matter what the domain of discourse, as Michael Dummett [1991] said, whatever is true is true "in virtue of something." But that is a plastic notion too. We have to distinguish "that in virtue of which" when we are talking about mathematics, logic, card games, and anything else you like, where the productive cause is our form of active thinking. In other cases, that in virtue of which our judgment is true is something other than the form of thinking, though the productive cause of truth is still our thinking. When "that in virtue of which" a judgment is true is an independent compliant reality, like the number of people on a bus, that is quite a different "in virtue of which" than, say, the invention by which I make up the features of a game or make an axiomatization of topology. The moral for now is: "whatever is true is true in virtue of something" requires decoding to fit the context and appears to offer explanation when none is provided.

(f) Truth does not require replicating things or matching things, or even things' fitting or working out; instead, truth starts with our wholly particular animal awareness and our coincident dematerialization of the components of judgment. That "start" is not an event but a continuous origin, even before birth. To understand is to be, for an intelligent animal. Cognition is not a match but an identification, for instance, I realize that someone does not understand what I am saying. What I realize is the very thing that is so. How that comes about might take a long story; but it is the only outcome that will hold up. When we know, we realize things, for instance, that "he does not believe me." In a true singular perceptual judgment our commitment is coincident with and has its content from animal awareness: "that guitar is still playing."

(g) Inquiries into truth have become separated from inquiries into animal cognition, and associated too closely with language, as if humans need not have had knowledge and true judgment in order to invent their languages. Truth, in non-cognitive systems -- whether encyclopedias or even sentences, is merely extrinsically attributional from the thinking expressed. There is no intrinsic truth or falsity to anything but a thought; all other truth or falsity is derivative or equivocal. So, if anything says anything, someone must have said it. Moreover, without understanding, there is no function for truth. We have to go back to the natural order of things, to rely upon the platform for judgment in animal awareness, including the presence of real things in animal perception.

(h) Judgment uses sensation the way gestures use movements, pictures use color or lines, and statements use words, as a medium; there are other media for thought as well, including bodily actions. Human experience has coincident animal and intelligent aspects, each involving distinct abilities. For humans, animal awareness, involving memory, imagination, desire, dreaming, perceiving, and feeling, is congruently conceptual and judgmental. Understanding is two constant activities, dematerialization and judgment (according to the operations "what?" and "is so?"), with a locked-on, but defeasible, reality commitment that is a continuous existential commitment. I will say more about these elements later. Additionally, there is the progression of judgment that we call reasoning. Reasoning supplies for an imperfection of human understanding, its incompleteness, and that imperfection gives rise to the "space" for logic and other formal systems, for you do not need transformations if everything is apparent to you.

Truth starts with our wholly particular animal awareness dematerialized and being focused on features within animal awareness. Conception is a constant activity along with our, as it were, causing the objects of animal awareness to emit light, to fluoresce, so that we can distinguish shapes, or color, or bones, or natures, or being alive, or objects, or "whats," as we happen to conceptualize. Certain native American peoples said everything has two forms, the outer behavior and appearance and the inner spirit-like form that makes the thing do and appear as it does. Analogously, physical things have appearances and looks, as objects of animal awareness, but concurrently transformed, they glow with structures, with generality for humans. We humans transform animal awareness, but we typically do that directly to things seen, heard, felt, smelled, moving and moved, etc. They have generality for us, though they get that generality by what we do and do not have it in themselves. As Avicenna said: intellectus fit universalitatem in re. Real things and their conditions, e.g., actual sunshine, are present in animal perception and present, conceptualized, in judgment. Abstraction as an operation on mere appearances is untypical and derivative, though we do it on sunrises, sunsets, the motion of the moon behind moving clouds, and trains passing in a station.

3. Now we have to dislodge some metaphors.

(a) The first: The matterless mirror.

The mind is not a matterless mirror, a scene for the mind's eye, displaying things in colors, sounds, feels, tastes and smells, and in shapes, sizes and dimensions produced on the mirror (i) by sensory transmission from the world, or (ii) by electrochemical sensory reaction to the world, or (iii) by some mind-initiated response to such reactions. The "inner scene" hypothesis is as if our experience were a Heads-Up-Holographic-Display, mostly computer-generated, by which we fly our planes over terrain we have no way of comparing, directly, to the display because we cannot see out. We have an inner display that we take to be close enough to outer appearance, and from which we infer outer conditions. That is "representationalism" sketched broadly, from which "the veil of perception" is instantly evident: if we do not know, independently, that the HHUD is designed to be reliable, how do we know things are as they seem? One obvious option was, as Berkeley did, to take the elements of the HHUD to be the real things themselves. Many varieties of phenomenalism were invented and elaborated only to founder because they could explain neither counterfactual truth (as required for what we might have seen/heard had we observed from a different position) nor law-likeness nor escape the inevitable nominalism they required of the same name for different sensory modalities (e.g., "square" as seen and "square" as felt).

(b) The second: The structure of the mind.

The mind is not a surface with its own reflective-geometry, like a lens, that patterns sense qualities into "whats," like mirrors in a Fun House making "thin man"/"fat man"/"sausage man"/"dwarf man" out of the light they refract--the patterns either (i) being the structure of "the mind-itself" (Kant) or (ii) arising transiently, from innate ideas -- whose fidelity, as a reflector and in each reflection, has to be justified. Ideas are, in such schemes, chiefly presentations of something else, thus, representations; so, there is a "veil of perception," and a general doubt about whether things are as we conceive them. (Of course, Kant's version in which phenomena are constituted from such structures and the existence of such structures is transcendentally deducible, is designed to escape the veil of perception, just as Berkeley's phenomenalism was).

(c) The third: The invisible subject.

There is no "fleshless eye," no "interior man," no spectator in the theater of inner experience, no homunculus, to "see," combine and "affirm" the presentations of sense, memory and imagination (the representations), thereby making judgments and reasonings that are true or false. There is no bodiless inner self "watching" an inner display, like a captured princess bemused by the forest/cliffs/sea-appearance that fills her whole tower window (which fills the whole field of vision). Dualist representationalism made an implicit two-selves story that doubles the problems without resolving any.

Further, no one, not even dedicated associationists like Hume, has ever come up with a natural-history of impression-combining that would amount to judgment. One cannot build assent entirely out of associations; for one thing, denial becomes inexplicable. Just as putting words together into a complete sentence, as a type-setter does, is not saying anything, so associating the impression of flame with the impression of heat is not believing anything, either. Neither is associating the verbal sound "devil" with a visual image of a horned, reddish, tailed male with a pitchfork, surrounded by flames.

Rather, judgment uses sensation the way gestures use movements, pictures use color or lines, and statements use words. Sensation is the basic medium, within which there are a variety of sensuous media, from images to actions, for thought. Still, understanding requires animal awareness (memory, imagination, desire, dreaming, perceiving, and the like); and understanding both makes and uses a constant, concurrent and coincident dematerialization of animal awareness resulting in conceptions, according to the operations, "what," and "how," with a "default," "is so," existential commitment.

The match-up (correspondence), fit-together (coherence) and work-out (pragmatic), notions of truth are all outcomes of the 17th century way of ideas, the representationalism that requires an account of representational accuracy, both individually (as images) and as complexes (judgments). As we will see, this was a completely manufactured problem, made up by the mistake that real things were not to be counted as constituents of animal perception or of human thought. It evaporates when sameness of perception and things perceived is properly premised and motivated.

Knowledge, from the 17th century on, was thought to require true believing (true judgment), where a true judgment is a faithful replication of reality, present to awareness and affirmed (assented) to be so (i.e., a faithful representation). Moreover, Descartes and all the major philosophers since, endorsed the "immediate awareness principle," that "what we are immediately aware of are our own ideas." That created the "veil of perception" and invited the more general "representational skepticism" (Hume's later formulation), to which the most prominent solutions were the phenomenalisms of Berkeley, Leibniz, Kant [Wilson, 1987 and 1995], 19th century idealisms that I bypass, and the twentieth century "scientific" phenomenalisms, like Russell's [1948], C.I. Lewis's [1946], and the latest versions of "materialistic phenomenalism" to be found in Carnap, Quine, Sellars, Goodman and Putnam. None of the phenomenalisms can surmount the technical objections concerning merely possible experience, counterfactuals, law-likeness, the impossibility of analyzing physical object statements into finite conjunctions of sense data statements (with resulting unverifiability), and more. More importantly, none can say anything informative about the origins and nature of animal cognition.

To reframe the traditional questions about truth, truth-gaps and cognitive access, I redeploy some positions defended elsewhere, most of which have been argued by other philosophers as well: that there are no empty names, no empty kinds, no merely possible individuals, no natures without cases (no empty universals or empty laws of nature], no subsistent abstracta, and no articulate domains of exemplar ideas or of other real abstracta. Physicalists should not want any of those things, either. We should, then, be able to agree that there are no truths about unrooted possibilities and no empty necessities with content. Thus, there will be semantically well-formed expressions that have no truth values, and others that do not satisfy bifurcation; consequently various discourses will not be suitable for "indirect proof," so the principles of validity will not be the same for all sorts of discourse. Michael Dummett [1991, p. 11] was, in my opinion, too hasty in saying the principles of valid reasoning cannot be a matter of taste. To an extent they can, for instance, depending on how luxuriant a mathematics you want to tolerate; and anyway, there can be different principles of validity for different sorts of discourse as he acknowledges: the principles are determined as much by usage as are the meanings of the words. In unbound discourse we do not use the principle that "a conditional with an impossible antecedent is true"; but sometimes we say "if he's a millionaire, I'm a monkey's uncle"; and there is the Irish phrase, "if he's a saint, there aren't any." And also, sometimes we say "impossibility of the antecedent yields falsity of the whole," and apply it to "if I'd been at Agincourt, I would have been king." What rules we use depends upon our discourse.

Human experience is not sensation or understanding or judgment presented to a subject, anymore than animal perception is. It is a subject's seeing, understanding, smelling, hearing, raging, loving, enjoying, etc. Just as for a plant, to be is to live, and to live is to feed, grow, reproduce, etc., so for a human to be is to "animal-rationalize." It is to do what a human does. The self, the subject, is the zoon logicon (and more). Whether there is a transcendental ego, lying beyond the horizon of experience that includes the experienced self, is not germane here.

An ever growing literature, from Rorty to Putnam, is converging toward the opinion that the projects of modern and of analytic philosophy failed, and indeed, some think the analytic movement rebounded to the idealism it began by repudiating. So it should not surprise us that the implicit dualism of materialism is gradually becoming obvious, not merely because abstract entities like numbers, logic, and laws are postulated, or because there is an uninvited "inner spectator" to every physicalist account of knowing, or even because crude behaviorism and mere information processing cannot explain the "display" element of the representationalism they are committed to, or the animal subjectivity they have to deny, but chiefly because the cognitive states are logically independent of the external physical states that cause them. It is a wry twist that physicalism, dedicated to repudiating the fleshless eye and all abstracta, becomes dualist in both respects. And, further, falls prey to the "veil of perception" in a way that parallels its archenemy, dualism. For it allows the possibility that we could be having all our present experiences when there is no "external" world at all or at least, none like what we experience. I take it to be Putnam's point in "Brains in a Vat" that such a supposition is absurd because the words of such brains would not mean the same as our words because the referents would either be lacking or different, and, thus, if such a situation were possible, meaning skepticism about our own discourse would be inevitable.

So I develop another standpoint, counseled from a larger span of history, to restrict the role of representations in cognition, chiefly to that of material-states-with-which, the way the chemical changes on film exposed through a camera are material states with-which an image is made on the film, though there are many functionally equivalent processes, just as there are several functionally equivalent processes of automatic focussing of a camera, too, and to say real things are continuously but not exclusively, or always prominently, the constituents of our thought. This is a return to realism. For example, it is your house I see, your car that I think to be blue; you I take to be here; you I understand or love -- not by inference from some image or representation that depicts or represents your being here, as if I were seeing your video-picture and thinking that you move gracefully -- but because what I see is what is. That viewpoint also counsels saying the content of our (and all animal) experience is naturally necessitated by our physical environment and our own species-specific physical constitution, and could not occur in its absence. Hypotheses to the contrary have no basis in observation or speculation, as I demonstrated above [Chapter 2]. To philosophize from such supposed possibilities is to build in mud.

I deny that what we are immediately aware of are our own ideas (things made by the mind) except in unusual cases. Even the cases you would think most obvious, say, pain or extreme pleasure, are not really cases of absorption in something we make that only represents the real, but rather cases of narrowly focussed perception of the real. To be entranced is not the same as to be absorbed in our fabrications or daydreams (though it can be), but more often to have awareness full of some real object, say, a person, or a song, or some real condition, say, an earache. Normally, what we are aware of are the components in the (underivative) situations we think of; but we can be totally absorbed, say, in stating our view, and not aware of the component gestures, words, surroundings or even the light. We can be so absorbed that we are not aware that we are self-aware. Another kind of thinking is speculation where an imagined or conjectured situation is the subject whose components are judged: e.g., we imagine tomorrow's confrontation with a tax auditor. That is quite different from the thinking-to-be that is habitual and has things and the features of things as components: the day to be gray; the tea to be cold; the hand to hurt or to be tangled in a seat belt. Such thinking, with imagined situations, invented images, and other creations of thought is parasitic upon our ability to think with real things and conditions as the constituents.

II. How Can Things Be Parts of Thoughts?

There are two ways: by representation and by transformation. Transformation is fundamental and has two modes, animal awareness and human judgment. We are concerned primarily with the transformation of animal awareness by which objects of animal awareness become components of judgment. But we will not wholly ignore the physical transformation by which an animal perceives its food, mates, habitat, etc. However, explaining how animals perceive things is "the great problem" I mentioned in the Preface. Representations, as parts of thoughts, are consequential, incidental and arbitrary; as parts of the psycho-physical system (brain states), they are a medium with-which but not, except in special cases, elements of awareness or judgment.

One is forced into representationalism, (taken as the view that the immediate objects of awareness are at most representations of the qualities of things) by (i) the immediate awareness principle, (ii) the principle that some sensations bear no resemblance to the distal features of objects that cause them, but are species-specific reactions, while others do resemble features of the distal causes (like shapes, sizes and distances), but not by replication or presentation of those features, but by surrogates that are like them. Those principles were generated by our supposing that all or almost all our cognitive states might be defective in the ways some can be: by illusion, dreaming, madness, deception, etc. We are then, like Descartes, supposed to find a firm footing from which to construct the non-defective cognitive states. Instead, I propose to deny the immediate awareness principle, and premise instead, that it is naturally necessary that our experiences have the distal causes they present to us, even though some of the features we perceive are species-specific as far as detection goes, and many features of things that may be sensible to other animals may be insensible (directly) to us: like the ultra-violet colors of flowers, to which bees are sensitive, or the electro-magnetic field disturbances by which some fish are aware of prey and predators, or high-pitched sounds audible to dogs but not to us. Moreover, I emphasize the question: what is the natural function for this sort of animal sensation (presentation, say, smells), as distinct from detection (litmus paper, smoke detector) in perception?

1. We do not live in representations.

We do not live in representations (copies, videos) of things, probing the world by habitual inference, like guards in an apartment house scanning floor-monitors. Neither do animals. Nor do we live in a permanent Holographic-Heads-Up-Display with innate graphics and sounds, etc. Representations, especially brain-state encodings may be necessary to explain perception and feeling but they are not sufficient and are not present to us.

Cartesian and Lockean representationalism are unconvincing; as Hume adroitly showed, they lead to skepticism. If all you can directly experience are simulacra, how can reason assure you that things are as science proposes? One way is by Descartes' appeal to divine veracity, a route almost no one would take now. Another way is idealism-phenomenalism: to say the material world is caused or constituted by ideas -- in effect, to deny representationalism by denying that there is anything replicated in the process of knowing but only presented. Phenomenalisms were attempted and foundered repeatedly, as I mentioned. Yet representa-tionalism keeps resurfacing, because it seems the only alternative to postulating magic and perhaps because it seems the best basis for making machines that think. Apparently machine cognition would require a device that converts inputs (that are detection-responses to features of things), first into a replication and then into appropriate actions. A mechanical player-piano is a simple detection-output device like a thermostatic furnace, though there is no "inner display." And, of course, there is no cognition either.

There is an ambiguity between "representation" as "replication of information" (as the perforations in the piano-roll are turned into mechanical leverage on the keys) and as "display that has to be 'read'," as J. Fodor pointed out. The Dretske information/transformation plan (e.g., an electronic player piano) is simply not complex enough to explain an amoeba, while the "display/read" plan implicitly reintroduces the inner seer or "little mind" to read the displays, something we can be pretty sure worms and computers do not have. Because neither disambiguation leads to a plausible plan, it is no surprise that machine cognition fares so badly, while robotics flourishes. The underlying notion of a cognitive system is defective and no amount of supplementation with computational and connectivist hypotheses will supply the key piece that is missing: the "display" requirement that demands subjectivity. If the perceiver has to read the display to experience the world, why not just have the perceiver read the world? That is my proposal. There is simply no way that the behavior of a squirrel or a trout can be just a software sequence of responses to environment, like a heliotrope or even the leaf curling of a hydrangea during drought. Now that does not mean cognitive science cannot succeed at elaborate information processing to do things we thought impossible to do without subjectivity (like simulating master chess playing with Big Blue), and also to make prosthetic-neural devices for humans and animals. But cognitive science cannot convert information processing into cognition.

It is as if philosophers had agreed that there are two main models: Either the material world, that includes our bodies, causes sensory states that are representations of other material states, or the material world is somehow constituted by conscious states. Looking at the bare bones of the options, neither seems remotely plausible, and most of the field of relevant options seems to be ignored. I follow a realist option in which there has to be a naturalistic account of animal consciousness of its world, and of the animal basis of human awareness, even though we have no more than promises from science so far, but an entirely different account of intelligent thought [Chapter 6].

2. Some remarks on representationalism.

Where do appearances come from? For Descartes, sensory impressions are not the "first state of awareness" caused by the last physical effect (the brain state). Descartes was particularly clear that ideas are furnished by the mind on the appropriate physical occasion, "by a special act of the mind" [see his Conversation with Burman], acting within a correspondence of physical states and soul-supplied qualities set up at entirely God's will, but for man's benefit. The mind's sensing is its experiencing a quality that interprets some brain state. Further, for Descartes, nothing but an intellectual soul senses, so that animals are just robots, without interiority.

There is something to be learned from this, to which we will return: that nothing physical is made-mental. We can also anticipate that whatever connection there is between brain states (very probably as types) and impressions is naturally necessary; and consequently, to suppose we might have had the same subjective states in the absence of our brain states (as Descartes supposed) or even in the absence of the necessary distal causes of such brain states (the physical world) is to suppose an impossibility. Now I do not mean that direct electrical or chemical stimulation of the brain cannot cause experiences that are indistinguishable from distally caused experiences; but in the absence of natural distal causes, brain states could not replicate or construct these experiences. That I might be a brain in a vat is simply impossible. And we can also note that animals must have interiority (what T. Nagel calls "subjectivity") though that does not require a focussed, reflexive awareness or a subjectivity capable of our feelings of hunger, grief, joy, etc. But it does require that a thing lacking "res cognitans" can be made of micro-particles so arranged that it is actively conscious and has its own (by type) subjectivity, e.g., "golden retriever-ness." Dogs and eagles can see. The contentedness of grooming chimps, and the sadness of a widowed elephant is observable from the behavior. That subjectivity can be very limited, as I conjecture it is with worms, mites and robot-like insects. We have to avoid claiming that some physical state turns into impressions; rather some physical states -- as part of a system of such states -- are the medium of impressions, perceptions, emotions and actions. And, apparently, there can be a vast variety among the brain-state media. Now how do any physical states get to be the medium of sensation?

The traditional representational accounts, mainly applied to humans, just beg the question. How would it explain a street-scene's being seen by me, to say there is a street scene/appearance, made up of minimal (threshold) units of appearance (something like an pointillist painting or a mosaic), of which units I am immediately aware and from which I infer (or somehow construct) the street-scene? First that entirely sidesteps the question of how the electro-chemical brain states (or fluctuations among them) produce the elements of the appearance. Secondly, the external perception problem is resolved by postulating an inner seer, hearer, etc., of impressions, for which (homunculus) there is no account, and from which the homunculus has to infer or construct, with doubtful accuracy, the external scene. How? There should thus be two experiences, one of the combined units and the other of external awareness along with an explanation of how the second is made out of the first. But we have only the latter.

Further, there is no way to explain how the units can appear to the homunculus. The "inner theater of experience" doubles the problems: who is in the theater having the experience, why does its experience have distinct but unified sensory modes, and how does that spectator know the inner theater presents the external reality? Would we really suppose there are fish-personae behind fish eyes, and mice-personae behind mice ears, etc.? Materialist representatonalists are implicit dualists, as David Braine claimed [1993]. Even Dennett backs into dualism by making intentionality of a system attributional; that is, a feature of a system resultant from what counts as the "best explanation of its behavior" by another (intentional) system -- an external homunculus. No gain. Physicalists, and materialists generally, are dualists, supposing that even animals have a theater of experience with an unmentioned inner subject, whose experience is logically independent of its distal causes, and, these, might have happened without them. Among idealists, the "subject of human experience" gradually became a "transcendental ego" lying beyond experience but "having" it; thus, a second kind of dualism came into fashion and is to be found, I think, even in Wittgenstein's Tractatus, as well as in later writers.

One thing a nondualist has to share with the most dedicated physicalist is that the animal's fears, rage, hunger, desire and hurt -- not all of them have all of those, of course -- by which they act in their econiches according to their kinds, are real. So also is their perception, whether by smell, taste, touch, sight, hearing or otherwise, real. How that can be must not remain unexplained.

Four key points in representationalist stories are inexplicable: (1) how brain states become (or are the medium for) representations that are displayed; (2) who the inner seer of the displayed representations is; (3) how the self, that has experiences that representations display for it, finds out that the world is as displayed; and (4) how the animal might have had the same experiences absent any distal causes at all.

Another difficulty with representational accounts of cognition is that representations are far more arbitrary than is usually suggested. People who draw badly readily recognize how arbitrary representation-relations may be, and how dependent upon internal contrasts among one another:

=happy?

(sad?) (person?) (learned person strutting?)

You encounter a form of skepticism: that, of course, the representations present something, but how closely does that replicate the world?

Some won't want to reconsider the direct realism options because, like Putnam [Nussbaum and A. Rorty, 1992], they can't see how to solve the problems they find, particularly about the status of real natures and the identity of the understanding with what is understood (though they don't seem to doubt that my piano playing can be the same as the scale played). Theirs are justified hesitancies. Yet, there is a lot to learn just from the fact that Aristotle's limitations on some points lie exactly where the Modern and Post-Modern philosophers have been stumped too.

For instance, Aristotle did not offer any explanation of how sensory appearances are made at all; he just postulated, on observation, that matter, the elements, can be arranged to form a sensitive substance. That seems obvious with birds, bees, fish, and animals generally, even though we cannot explain how it is accomplished. In fact, in De Sensu et Sensibile [1984 edition, 438(a), 13], arguing that the eye is transparent, Aristotle seems to think the appearances of things (the color) passes through the colorless eye on inside us so that the sense ability becomes the color. In another passage from De Anima Aristotle seems to say the sense organ is not colored by the color it responds to, but the actuality of the sense-power is the actuality of the sensible object qua thus sensible, that is, the actuality of the color-sensory power is the color of the object (that is what I think Aristotle means). But Aristotle does not provide a physiology for such a process or even show how there could be any physiology to explain it. And, on that point, he was stumped right where we still are.

Still, Aristotle makes persuasive that the sensory consciousness of humans is analogous to that of other animals. The reason it is analogous, rather than exactly the same, is that in humans there is never "just" animal awareness; it is always suffused with understanding (even if very limited or infantile) because the intellectus agens, the "making intellect" is always "on." Animals perceive, not just sense; animal cognition is a continuing part of human cognition, though transformed by intelligence. Understanding is a constant transformation of animal awareness (for humans); it is not understanding that makes perception because unintelligent animals can perceive; rather, intelligence makes perceptual judgments that are true or false. Animals do something like judging; they perceive, but not with content that is true or false but only accurate or erroneous (measured by payoff). And, of course, as I elaborate that basic position, it also turns out that the vital principle in humans that is the principle of both intelligence and sentience is not a material principle, as it is in animals. It is not material software at all.

3. Internal to external awareness. No.

Rationalist and empiricist accounts of human knowledge have another defect. They explain knowing as if we progress from internal objects of cognition -- our ideas and impressions -- to external objects and situations. In fact even recent and analytically acute accounts, e.g., R. Chisholm's accounts of the immediately evident, follow that order. So do Lockean representationalism and all the forms of sense-data phenomenalism. But, go back a step. Suppose we do not (individually, or philogenetically) develop from a state of "internal" awareness (say, of the immediately evident, the phenomenal) to "external" awareness, say, from a state of impression-awareness to a state of object awareness. Suppose, instead, we mature individually from chaotic infantile "well-being" awareness, unpolarized, neither internal nor external, with motor responses (cries, movements), propelled by contentment, hunger, pain, rage and fear and abandonment, that gradually become refined external awareness of people, objects, conditions and ourselves, that are in a dream-like externality that gradually "clears" of haziness to crisp external consciousness. Suppose we are eventually or concomitantly socialized to internal "awareness and centeredness," to a self-awareness organized by a self-conception that we learn, and suppose we mature as we grow toward adolescence by "knitting together islands of consciousness" [cf. C. Jung], so that portions of our experience are not dissociated from one another -- though for most people there will always be gaps, sometimes surprising ones. Suppose inner life is not the progenitor of the external life, but is its offspring. Suppose only higher animals have "knitted together islands of consciousness" amounting to a continent of organized self-awareness that includes will directed to what we understand (personhood); suppose that other animals, to various degrees, have blank unconnection between hunger and mating, or imagining and acting (though they may be related and ordered by instinct). Then we can start the explanation of knowledge in a different place: Externality is first, even for intelligent animals.

4. Aristotle and Descartes: Nothing physical becomes mental.

It's an old principle found in Plato, employed by Aristotle and Descartes, that no material cause has an immaterial effect. Formulated more particularly, nothing physical becomes mental. Consciousness, even animal consciousness is not a condition of material being, qua material, otherwise all of matter would be conscious. So, philosophy that claims there is only undifferentiated basic matter under one set of universal organizing laws and no inherent structures to things, no specialized software in things, has to be mistaken because all of matter is not alive, nor is it aware. Another step is needed in this reasoning because not all matter is machinery either. But no one doubts, now, that classical mechanical principles are not the whole of basic physics. So, why not a supplemented physics as the explanation of life and sensation? Especially physics that includes chemistry? The organization by which some matter is alive is as real (and variegated) as is the arrangements by which field-stones can be made into walls; in fact, there seem to be between 100,000 and two million basic molecular arrangements sufficient for a living thing. There is no single software for life. Nor is there for sensation, perception or animal action. To say that the arrangement or organization that uses the obediential ability of the components is not as real as the components and their passive capacities is to leave a complete pile of radio parts and a functioning radio without an explanatory difference or a pile of squared stones and a buttress or arch as no different. Whether the organization is reducible to or a consequence from universal physical law has nothing to do with its reality in the organized thing. The organization of the stones is just as real and effective as the buttress.

Nevertheless, the principle that nothing physical becomes mental operates differently for Aristotle than for Augustine (sentire non est corporis) and Descartes. For Aristotle sensation, feeling, memory, imagination, imperfect will and action (a kind of agency) are within the ability of some physical things, animals, and do not originate in an immaterial soul, whereas both Augustine and Descartes thought action, sensation, perception, and feeling, like understanding, originate only from an immaterial, intelligent soul.

Moreover we get nowhere, to say that the very same thing that is the last physical effect, a brain state, is also the first thought-unit, a sensation, without explaining the cause of (and the quality of) the subjectivity. Whether the last effect is an electro-chemical charge-pattern or a fluctuation (function) among them makes no difference: what makes it "a content of sensory awareness," something felt, seen, heard, smelled, tasted, a touch, or feeling? Is there emergence? It was never widely believed that something has the effect of turning the physical into the mental. Instead, there was prevalent Augustinean/Cartesian talk of "the mind's attending to" the states of the body, and Descartes talks of the soul's attending to states of the brain -- what I have called "scanning" the body, apparently by some super-awareness that precedes, and is not, sensation or perception. But that tradition just seems to repeat what is to be explained. For such talk supposes the body and its states appear as contents within the scanning mind, like parts of a mountain within the scanning binoculars. If the mind provides the contents of the scan (as Descartes says it does because all ideas are innate), then, strictly, how is it "attending to the states of the body"? By producing the sensory correlates (by divine decree) of the bodily condition? But how does it know when to produce a correlate? And, thus, how can the soul know the bodily condition? The soul's loving vigilance seems vacant. If it could know the bodily condition directly, there would be no need to produce a correlated impression. Failure of that sort of account invited the psycho-physical parallelism (ideas/material mode identity) of Spinoza and the pre-established harmony of Leibniz, though neither of those views explains the presence of ideas to the soul.

Aristotle did not succeed either. Abstraction from phantasms (unified sensory appearances that contain forms), might explain conception; but no explanation was offered for how the cause (a bell's ringing) of the unappearing stimuli (say, sound waves) becomes an appearance (sound) which is the same as the condition of the active bell heard, except to postulate an organ whose function is to hear and whose response to the signal is the same as what causally originated the signal. Some of the late medieval Aristoteleans developed theories as to how the form of the physical trait (light) could be transmitted materially without informing the medium, to become the form of the sense power [Simmons, 1994]. Sound and smell do, of course, inform the atmosphere. Hearing is not like a radio broadcast of a concert, where the sound is turned into the broadcast signal, received as radio waves, amplified, converted into magnetic impulses and reproduced (on speakers) as sound. For the sound heard in that case is not the original sound but another sound, as much like the first as technology can make it. That is not Aristotle's idea: the sound perceived is the same sound the bell makes. So, too, the color perceived is the color of the brick, the smell is the smell of the rose.

Aristotle may have been on to something. The telephone presents mechanically the same structure as a radio, though the transmittal signal is low voltage electricity, except that because of the phenomenon George Mavrodes [1970] called "input alignment," we can hear and talk to the person calling even though the sound by which we do it is a high-quality effect of the sound the speaker is making. Yet the statements, questions and feelings we hear, via high quality replicas of the speaker's sounds are what the speaker says, etc. So Aristotle's basic idea is one we can illustrate in human perception and understanding. Nevertheless, it seems to me now that perception is not done by a reconstitution of the features by which something is perceived, not even by digitally perfect reconstitution that would leave no physical difference between the originating quality and the qualities sensed. For that would still leave us with two kinds of hearing: the hearing that senses sounds and the hearing that hears things, bells, voices and people. No, the best account, even if it deviates from Aristotle's intention is that, normally, the animal perceives things or the features of things. How the brain accomplishes that is, admittedly, a mystery so far. But how the brain accomplishes any reconstitution, as described above, is an even greater mystery because it requires an additional awareness of sensation besides awareness by sensation.

Aristotle's idea, I think, is that a sensory system is one whose response to the transmitting signal (e.g., to certain shaped molecules fitting into places in the nasal system and triggering a neural-brain response electro-chemically, as we describe smell today) is the feature transmitted: the original smell and not just an effect of it as the red color of blue litmus paper is an effect of acid, or even a perfect reconstitution of the original. That is, it is not by way of input alignment, as we can see the moon through a reflecting telescope presenting a moon-image on the mirror [Mavrodes, 1970], that we see a person. Rather a sensory ability is made actually to sense (hear, see, smell, taste, feel with a certain content) by the very feature whose signal-effect is transmitted through the medium physically to affect the sense organ including the neural system and what we now know are the responsive areas of the brain. The sense organ undergoes physical changes from the signal, but sensation is the activity whose content is the feature of the object sensed that distinguishes a sense from a detecting device (light meter or litmus paper) whose response is caused by a feature of things that is neither the same as, nor is any mode of experience. Now sensation is not ipso facto perception; organization is needed for perception, even for perception that falls below direct attention. Thus perception is complex sensation that instinctively (and from learning) presents something to the animal, eliciting action that is proprioceptive as well. That was the idea. But how such a system can actually exist physically has to be explained and so far, cannot be, though, as my telephone example illustrates, we certainly do experience the very statements and questions and the real presence of another person, and actually have heard the person speaking to us. There really is not that much difference, as far as perception goes, between seeing something by way of a mirror and seeing it by way of a video camera. But that is unlike visual perception generally because both cases require one kind of seeing (the image) for seeing the other (the original). The right account of normal perception is not by input alignment of a feature transmitted by a signal and reconstituted as a sensation. Rather, the perception is a mode of direct awareness.

Moreover, all other accounts of sensation that proceed by causation of a signal from a distal source stimulating the signal-sensitive organ and neural path to a sensitive brain sector, end in the proposal that an inner sensory effect, like the voice-sound on a telephone, is produced in the animal and that in response to that, the animal detects a distal feature of the thing, apparently having learned to recognize things from their appearances. But that adds a stage to what the Aristotelians could not explain: for now we have the "presentation of a representation to awareness," inner seeings, hearings, etc., and some kind of process by which the animal adjusts its actions to its impressions, as if it were an HHUD-system: two more physically unexplainable steps. And the inner subject now has a new task: to reason (or act by instinct in animals?) from the HHUD display to the features of the world that may not even resemble the features of the display [cf. Descartes, quoted above; Sellars, 1963; P.M. Churchland, 1979 and 1995]. There is even less explanation of how to realize that situation physically than there is for the my supplemented Aristotelian view. In fact whatever physics/biology/neurology would explain the latter representationalist story can be adapted to explain the direct perception story. After all, if I can hear you when the sound I hear is a reconstruction by magnetic impulses of the sound you make, then in principle, if you were in the room, I could hear you directly.

Aristotle as I present and supplement him may appear to require magic; but the representationalists require magic twice: (i) about conscious response from a physical change in the brain and (ii) about how the animal gets from the subjective to the objective world. The only magic required for animal perception on my account is how the brain can digitally present the very qualities of the object.

The heart of the issue, which also divides the interpreters of Aristotle, is whether (1) sensation is by a transformation from physical signal to sensing response, where the originating physical reality is replicated (duplicated = species) in a sensory response (phantasm), though the real object or its quality is what is perceived, or (2) whether the sensation that is the response is the very feature of the thing (present intentionally) that initiates the chain of causation. I favor the latter, regardless of which is the better reading of Aristotle. In either case, the higher animal's perceived world is considerably like ours, absent all the "whatness" of it, and adjusted species-wise as to what features are perceived in things (e.g., ultraviolet color for bees and black and white from low-angled wide-scan for the visual world of dogs). Another limitation of that general outline, in addition to the mystery as to how to realize it physically, is that it does not explain exactly how the eye presents colored things, rather than just detects them, the way a smoke detector detects but does not smell smoke. For it is the intention of the Aristoteleans from the beginning that perception is awareness of the real features of things (colors, tastes, textures, odors and sounds) and though them, of the common sensibles (shape, size, weight, etc.) and through them, of things as behavioral entities, substances and, for humans, coincident pre-scientific grasp of the natures of things (e.g., of potatoes vs. tomatoes).

Detection, unlike sensation, produces no cognition unless attached to an interpreter, and usually, like a thermostat, has no cognitive role, but only serves in a system, that is arranged to start, stop and modulate activities. Thus the automatic landing system on airplanes links altitudes speed/pitch, attitude, and other detectors through a computer to position the hydraulic systems that control the plane and, by constant feedback, to land it. The entire system is just an elaborate version of an automatic pilot for a sailboat, the whole being a servo-mechanism with a target. Sensation is quite unlike that. Its function is to contribute to a species-specific presence of other things which elicits instinctive (and learned) action. The intervening presence (which can be very primitive) is what needs explaining, even admitting borderline cases (e.g., one-celled animals and certain plants that do not sense, or if they do, do not perceive). What is the function of presence and feeling in nature, that could not be performed by detection systems without subjectivity?

For Aristotelians, the function of sensation is perception which in turn has as its function action for animal-well-being. Sensation is the means of perception which is the presence of things to animals. Animals don't just detect predators and prey; they see them, hear, smell, taste, and touch them. But the presence of things-understood to humans requires an activity peculiar to humans, not using a material organ but transforming the product of material organs, phases of animal consciousness, into universality (conception) that in the atmosphere of desire and consent, prompts judgment. The basic activity is dematerialization by which form is separated from its matter (like fluorescence that makes bone emit white light) so as to be content for understanding. Aristotle explicitly describes abstraction as an operation upon the phantasm [De An. III and Aquinas' Summa Theologica, I, q.84, a.6], that is, upon the sensus communis-product of animal sensation. On my account, the normal object of abstraction is the thing-perceived, not its appearance alone. So animal cognition is the base, the platform, for human understanding. In fact, without a sensory medium there can be no new experience at all. Normally it is by appearances that we perceive things and only rarely do we perceive the appearances; but we can do that with an effort if we want and by happenstance when we realize we seem to be moving but really are not because something else is.

Lest the idea of internested emergent systems seem a mere promise, here are some brief observations. First, there is a simple notion of active form: take stones, not actually able to pile themselves into a stable wall, and assemble them according to their balance, shape, angle, etc., into patterns where they hold one another in place, form a straight line, slope inward from the bottom, etc., where no forces beyond general physical forces (gravity, push and heave from frost, vibration, from traffic, freezing water, etc.), operate, but where the arrangement, which is obviously real and from the hands of the wall-builder, orders those parts and takes advantage of all those forces to hold the stones in place, the sides vertical and the top level and all the supporting courses in place. The complex arrangement is an active form, so complicated we can only describe it schematically yet all it does is channel otherwise existing forces.

A more elaborate version of such real structures is the Roman aqueducts and the Roman highways that stand after 2,000 years. That is structure entirely realized materially, but involving no active forces other than physical, gravitational, mechanical (and in some cases hydraulic) forces, but ordering the effects of such forces into structures that have long-term stability, etc. We cannot analyze the success of such building without notions like "thrust," "balance," "support," all active characterizations of what the design does. Design, in such structures, and in buildings and in ships, is real, intrinsic and active form. It is entirely material.

I will not go into other kinds of arrangements, such as inner mechanisms in mechanical clocks, internal combustion engines, a toilet, a stone sewer system, the structure of buildings and the like until Chapter 7, beyond noting that they are all examples of active design. There are many more elaborate real designs and forms that are arrangements of the material propensities of things to achieve certain effects, some of which come about by the general course of nature, like tectonic plate movement, glacier movement, river bends, mountain ranges, and many others by intellectual fashioning, the way flint knives and arrow tips were made, as well as rifles, assault weapons and drones.

That loose family of forms -- all involving organization of "passive" elements under general natural forces -- contrasts with various other kinds of composition in which the forces operate differently, in effect, under regularities they could not initiate on their own. That is the basic notion of an emergent-system: regularity of behavior achieved by structurally ordering the active natural capacities of things toward effects attributable to the structural ordering, which the separate capacities could not achieve on their own. The evolution of vision and hearing are examples. A forest fire has to be ignited by chance or intent or necessity (from an adjacent one), but what happens after ignition is determined by a mixture of the active and passive dispositions of components, prevailing winds, moisture, etc., until a level of ferocity is reached where the individual components and large groups of them are irrelevant to the fury. The whole forest fire forms a unity that progresses under laws that are not laws of the components, as is also true of ocean waves.

Such, in rough outline to be refined later, are "emergent systems," where the system behavior is beyond the active ability of the components, and where the active and passive abilities of the components are "ordered" to behavioral outcomes that, however, employ no forces other than universal physical forces, as particularized to mechanical, hydraulic, heat, electrical, magnetic, wind, moisture, gravity, atmospheric disturbance, and the like. So the emergent form is entirely in the successful arrangement of natural forces in appropriate materials to achieve the outcome (a forest fire, or a larva flow, or a volcanic eruption). Bending sharp metal strips and hinging them is enough for a scissors.

That contrasts dramatically with the second phylum of emergent systems: where the system-behavior is beyond the active and passive abilities of the components, for instance, sensation or activity from desire.

Now I suspect that the explanatory principles of plant and animal life belong to the universal causal principles of nature but are conceptually underivable from the ideally perfect universal physics, even if we could ever get one. Thus, even though the ultimate components of animals and plants are all nothing but cases of the universal particles, fields and forces, the organizing principles by which they can be arranged into individual living and sentient things are, in principle, underivable from the universal cosmic physics, though they are principles of physical organization just as the jet stream laws and atmospheric laws are as well. Thus, I consider such designs to be emergent, rather than resultant or consequential.

For now, distinguish the subject of explaining animal cognition from that of explaining intelligent consciousness because the latter involves abstraction that is coincident with animal awareness and, if Chapter 6 is right, cannot be physically realized.

Physicalists are right up to a point: sensation, which is definitive of animals, has to be a feature of suitably organized material systems: of birds and dogs. You do not need an immaterial soul or mind for sensation, or perception. But physicalists are wrong to suppose there is no subjectivity peculiar to the various animal species and fail, thereby, to carry their share of the burden: to show how perceptual subjectivity and the perceptual presence of real things can be physically realized and why it is an evolutionary outcome, instead of mere detection systems. We need something more than the universal principles of physical organization, and, as we know now, more than the electro-chemical arrangement of matter which occurs in a radio or camera. That is, the system, or as I call it "software" realized physically, has to be intrinsic to the animal (not extrinsic as in printing or in a painting) and has to be more than the general principles of physics and electronics and chemistry and biology, yet realizable in the physical-chemical-biological systems as operating systems, just as WordPerfect and Lotus Spreadsheet can be realized in DOS, or OS2 or Unix, which in turn require certain machine languages and eventually, certain physical hardware and electron behavior, and in the end, the same physical particles, forces and fields that make up the cosmos. The principles of percipient organization have to be additional to the universal principles of matter and of plant life because only some physical things are sentient, percipient, or even living things. As Aristotle said, "For living things it is living that is existing" [De An., II, 415.b, 14].

There seems to be no option but for all disputants to agree that the physical realization of percipient subjectivity and action has to be explained, even if that requires a whole new inquiry into emergent systems, or even an evolution in the notion of scientific explanation. Secondly, we are not trying to figure out how in nature matter organizes to produce an inner-theater of experience in a physical thing, but rather an "outer-experience" of other things and of itself with a species-peculiar subjectivity to it.

If we eliminate two defeating models for dealing with the mental, we can make it pretty obvious, first that other animals, too, have to have real things as components of their awareness, and that human thought begins the same way, but involves additional activities that eventuate in judgments for whose truth and falsity we need explanations. We can then evaluate standard theories of truth and prepare for another one.

III. Two Defeating Models

1. First defeating model.

It has been proposed that the last cerebral state (in the chain of causation from a distal stimulus) is a sensation. We know it cannot be a token-token identity, because other people have the same sensations. It cannot be a human type-token identity because things of other kinds have the same sensations (both cats and dogs have "milk-taste"). And in general we can conclude there are resembling subjective states in many species from the anatomical and functional similarity of the perceptual organs, and from their differences, too, e.g., the different construction of the eyes of certain fish that see above and below water, and birds that have a telescopic center of otherwise wide-vision, with the result that there is not one type organ-cerebral state that is identical with any given sensation-type.

Functionalism was supposed to get around such limitations because one could suppose the same function among widely variant cerebral states, just as one function "auto focus" is done in cameras by three different physical systems. But functionalism has turned out to be empty. Although the hypothesis would escape the token and type identity problems, input-output correlations cannot determine the content of such a function tightly enough to determine a perception, sensation, or thought, any more than a mapping of inputs to outputs can determine a mathematical function uniquely [see Chapter 6 for the argument and the consequences].

In one way, content-failure may be the best news in a long time. For the "content" of the function is the story that gives the best explanation of the succession of input (which is determined by organ stimuli) to output (in the animal's behavior). That is, the precise content of the "inner state of awareness" cannot be uniquely determined by any input (physical stimulus) to output (animal action) "best explanation," because there are always competitors, mutually incompatible. The result is that the content of the subjective state is not determined by the best explanation of the causal regularity of input (stimulus) to output (action), and thus, the content of the inner state cannot be the same as "the" function of input to output, for there is no such thing as "the" function. So identity of brain state (or functions among them) with the subjective sensory states of animals turns out to be a defeating model of the mental-state relationship. That at least draws attention back to our main objective: to explain how an animal (without an immaterial soul) can experience itself and other real things and actions.

Many philosophers by varying paths got to saying matter can't cause knowledge, and that the mind must either transform the material to make it intelligible (Aristotle) or somehow constitute the objects (Berkeley, Leibniz, Kant, and later Idealism and phenomenalism) or get its objects elsewhere (Plato's recollection, Augustine's illumination, and Descartes' innatism). But notice, the Aristotelians did not regard animal cognition as from an immaterial soul, but from a material form destroyed with the animal. And they thought there is animal cognition.

Later medieval Aristotelians who followed Aquinas on the unity of the substantial form in a human (denying the Bonaventurian proposal of an addition forma corporeatatis) did hold that even the animal portion of human consciousness is from an immaterial indestructible principle, as is the being of the composite itself, though because the matter is corruptible, so are the animal powers. The human is able to sense by the immaterial soul but can be made unable to sense by indisposition of the flesh. They did not regard animal cognition, feeling and volition as "mental" in the sense in which Augustine, Descartes and Berkeley did. That first impediment that supposed all mentality to require an immaterial soul, can simply be ignored now. We can re-instate an Aristotelian realism that animal cognition is from a material design (form), but that means we have to take seriously the idea that there are real, causally active structures (forms) in nature [the point of Chapter 7], that are the suitable objects of intelligent understanding [Chapter 5].

A second impediment from the "way of ideas" is the corollary that cognition proceeds from subjective awareness (of sensation, feeling and volition) to outer, external awareness of things and action, (as the section "Internal to External" above observed); we reject that because its explanatory gap, "the veil of perception," invites skepticism. Instead, we ought to give more weight to our limited observations that suggest that in higher animals (horses, apes, dogs and birds), sensation configures the external (e.g., the bodily and the surrounding), but not qua external, because it is not presented in contrast with anything. The species-determinate subjectivity of animals is not in what they focus on internally, but is the sensory and emotional modality by which the external, including their own bodies, is presented and manipulated.

Understanding requires our doing something to whatever perception presents. What we do is treated separately [Chapter 5]. For that to happen, something makes the physical appear, by colors, sounds, shapes, smells, feels, with shape, weight, velocity, location and distance (not as we see or smell them consciously with concomitant conception), but as they can be to us when entranced, and are in animals. It is the latter that matter is passively capable of and in suitable conditions, emergently able to achieve. Human awareness adds a constant transformation of animal awareness by coincident removal of particularity in one or another respect -- what I call dematerialization -- with preservation of particular awareness.

The best working stance is that the "last physical" effect in the appropriate cerebral area is not identical with or turned into a sensory state, but that the entire system of brain states at the appropriate level, is material-in-which a further real material structure (which organizes the whole animal body and makes it alive as well as sentient) is realized, the way WordPerfect can be realized electronically in a computer. In animals, the specific form has to be intrinsic and throughout the whole animal right down to the cells. "Being a rabbit" is a form, an organization by which certain organic matter is made into a rabbit; it is what disappears with death.

In addition to the explanatory relation having to hold among systems (for defects in a brain area defeat sensation or action of that kind, unless the activity is reassigned), the explanatory order has to parallel that between meaning and marks: the explanation has to go downward from the perception and action of the whole animal to its brain states (or fluctuations among them) and behavior. For an animal, perception and action as a whole has to explain the cerebral function as a whole. Thus the principle has to be a material structure, a form, of which the physical elements are obedientially capable, the way silicon and metal parts can be made into a radio. It is as if there were a formula that the animal cell-parts have to satisfy the way there can be a radio-plan that the parts have to satisfy for the radio to function. Perception is for the animal; the brain and organ states are for the perception.

When we reject the "way of ideas" and the resulting "veil of perception" for animals, and say sensations are not produced by an immaterial soul and that animal perception and action is not the same as an internal cerebral state and that the external causes are naturally necessary for the animal perceptions and actions (so that animals could not have the same experiences absent their material external causes), then we are ready to recognize that even for animals, real things are components of animal awareness. That provides a fluid transition to the idea that natural human cognition is always a transformation (by dematerialization) of the platform animal cognition and action.

2. Second defeating model: Attributional-intentionality.

This can be disposed of quickly since it has already been described. D. Dennett [1978] said the intentionality of a system is equivalent to whether the best explanation of its behavior requires ascription of mental states to it. Hydrangeas wilt in hot weather to reduce the area exposed to the sun's evaporation, so they can survive long heat and drought. We might say, "the hydrangea knows how to conserve its water supply by reducing the evaporation area of its leaves." This is an example of treating a material system as an intentional system, with attributed intentionality. But that attribution is not required for the best explanation. So we do not say literally that the plant knows what to do. Similarly, we use metaphors in talking about computers: "it stores it," "it recalls it," "it spells it right." We say of cars, "that car wants to pull to the right." We say the stock market reacted with fear to rising prices of consumer goods. These are metaphors for easy description but not part of the best explanation, so no mentality is literally attributed. But with humans and animals we do attribute mental states, both at the thought-level for humans and at the sensation/perception and limited-volition and action level for animals: e.g., The dog wants to go out; the dog won't walk that way. It is a long and unjustified reductive step, of course, for Dennett [1978 and 1987] to say that such an attribution's being needed for the best explanation of the behavior is what, and only what, the mentality consists in. That also revises of what we mean, and thus commits a key error in philosophy: telling people who know the language that they really mean something other than what they think they mean.

Sometimes we talk mentalistically of groups, "the crowd panicked," "the stock market tried again to maintain the week's averages", "the cars moved so close together they forced the average speed down"; and we even attribute common consciousness to groups "the mob went wild, tried to trample the soldiers, retreated a few blocks, reorganized and rushed to burn the palace." The fact that we attribute intentional states is not sufficient for mentality; also required by Dennett is that such attributions be needed for the best explanation of the behavior. But even where that condition is satisfied, we have justified only the attribution of mentality, not the predication of it. The problem with the view comes, not because unnatural entities, e.g., utility power systems, the flow of traffic, and the Dow Jones average would as well as animals have to be regarded as intentional, bad as that is, but from something else.

The rock bottom objection is that the ascriptive-intentionality of systems, say of a utility's power distribution system, conceptually presupposes that there are systems that are inherently intentional. But such notions of (non-attributive) intentionality have not been explained or allowed for by Dennett at all. In fact, they were supposed to be displaced. But some things have to think, really (and non-denominatively), in order for the description of other things, as if they think (the attribution of intentionality), to be part of the best explanation of their behavior. For if nothing really thought at all, then descriptions of things as thinking systems would not be part of the best explanation of any behavior at all. In fact, such descriptions would not exist. The form of Dennett's explanation precludes its own success.

How scientifically to explain animal consciousness requires a reconception. The problem has to be formulated as to how to design a system that responds to its environment with actions it originates, for the success of which a presentation of parts of its environment and of itself is indispensable. There is obviously no advantage in an automatic sprinkler system's smelling, rather than detecting smoke. What is the advantage of smell to dogs? Some animals get by with sensation without perception; the ones in classes close to mammals have perception as well. For them, what is presented is presented as a feature of a distant or ambient object, or of the animal itself.

3. Now what can we conclude from all this?

First, that although the models and metaphors we looked at are faulty and misleading, still, the explanation of animal cognition is largely a matter for physical and biological sciences, since sensory cognition has to be a power (very probably an emergent one) of physical systems. But the manner of explanation will have to be in terms of real, dynamic material structures (forms) that can be physically realized, and can organize organic matter "from the top down" into percipient entities, aware of other real entities and of some of their own bodily parts and states, not by replication, simulation, simulacra or representation.

Secondly, the "way of ideas" will have no function in our explaining basic cognition, whether for humans or animals. Just the opposite at the basic level of cognition, the constituents of awareness have to be real things and their conditions, without generality for other animals and concomitantly particular and generalized for humans. Rejecting parallelisms of the mental (whether of thought, ideas, sensation or volition) and of an external reality for animal cognition, I can now turn to specific reasons for rejecting parallelisms of judgment and reality to account for truth. That will include any sort of "matchup," "work out," "hang-together" account of human truth offered as a global account of truth. Thus, I do not begin by asking whether bifurcation for all truth-bearers is to be supposed, or whether there is one or more cognitive element(s) in the notion of truth (like cognitive accessibility, warranted assertability, in principle verifiability, pragmatic success, fulfillment of expectation, etc.) or even whether truth, globally, requires "a fact of the matter" and whether such "facts of the matter" have to be independent of our conceptions and judgments.

One can already anticipate from the plasticity of the notion "true" that the phenomena are an analogous multitude, with local phenomena exhibiting the above mentioned features variously according to the discourse. As a result, one cannot follow the recipe that M. Dummett sketched [1991] for determining from the logic applicable in various sectors of discourse whether the phenomena purportedly discussed are real or not. For he did not make it persuasive that non-classical logic assures any irrealism at all. Nor can we accept Crispin Wright's proposal [1992] that there is a "core notion of truth" with accidental variants by context. Instead, we have to acknowledge that "true" means different things and has different conditions of applicability in diverse discourses [see Wiggins, 1980, for similar claims], and then concentrate upon providing some real content for what otherwise comes out to be a truism, but the only survivor for a global notion: that what I think is true just in case what I think is what is so.

IV. Bad Global Theories of Truth

In sum: The worst is a match-up, correspondence theory, because it is naively attractive yet no match will explain truth, as we will see. "Coherence" and "pragmatic" theories are sometimes locally useful, but are circular and self-destructive as global theories. The others, like the emotive, redundancy, and the disquotational theories do nothing to explain truth or falsity and cannot accommodate a large part of our usage of the notions. Does anyone seriously think that none of our judgments would be true if we simply dispensed with any metajudgmental predicate like "true" except where we need it in indirect discourse? I want to explain the relationship between my judgment that it is raining and it's actually raining, without any concern for the word "true."

1. No correspondence.

No correspondence of pairs with different components assures or explains any kind of truth: not sentence-parts corresponding with fact-parts; not thought-parts corresponding with reality-parts. Correspondence of pairs of things, like truthbearing sentences with truthmaking facts, is not like meshing the gears so as to make the sentence true. Such a "meshing the gears" is just what no philosopher has been able to describe. Similarly, to say some thoughts, statements or even sentences are somehow copies of pieces of reality, and thereby true, has turned out an empty promise: no one can show how any such thing could be a copy of reality and, even if it could be, why that would make it true.

Philosophers recently tried to explain the truth of what we say by way of features of what we say it with, treating "is true" as a semantic notion. Sometimes it is sentences whose assembly is supposed to key into the world. Sometimes it is abstractions smoked out of utterances: namely, statements, or propositions, whose internal assembly is supposed to match or fit the facts. They all meet the same fate. What makes what we say true or false is not features of the expression of judgment, or of abstractions from it, like propositions, but thought.

Part by part metaphysics, like "scale modelings," "mechanical drawings," or "pictures" [Wittgenstein, 1922] will not explain any truth: not formal truth, natural necessities, empirical generalizations, empirical conclusions, empirical inferences, abstract generalizations or simple observations. Basically there are two reasons.

The double decomposition into ultimate component parts, parts of reality and parts of thought (or parts of sentences) pair by pair, has never been delivered. B. Russell and L. Wittgenstein proposed such parallel decompositions for atomic propositions and atomic facts, so that for Wittgenstein the true atomic proposition is a picture of an atomic fact. But their ontic accounts of the "parts" were unconvincing. Besides, a parallel decomposition into ultimate semantic components is not even coherent. Unless there is a universal semantics for all 3,000+ extant natural languages, why should some ultimate analysis of "I have two jobs" in each, match up with the "ultimate ontological components" of the situation? Yet if the match is not of ultimate decompositions, what matches what? Had the "semantic marker" idea panned out and held up throughout contextual adaptation that causes polsemy, one side of the decomposition might have been plausible. But the ontic decomposition would fail anyway. Russell's "complete complexes of compresence" was a disappointing candidate for the latter.

No match-up, one by one, of parts of a truthbearer and compliant reality will make a truthbearer true, anyway. The parallel assemblies would have to gear together to do something we could see to be "being true." The picture theory that Wittgenstein suggested was supposed to be a way of being the same. We were supposed to see that "being a picture of" is "being true," especially by noticing how a mechanical drawing displays a machine, or an accident-sketch shows the scene. Our not being able to say what picturing consists in was supposed to be an advantage, by blocking a further question as to why picturing is truthmaking. We were just supposed to see that. But now we know that in several senses "being a picture of" involves no "truth" at all (e.g., see Goodman [1968], on "representation"). We know that picturing always involves construal and that there can be endlessly variant pictures of the same situation and pictures made up to show what never happened at all. Similarly, I think Wittgenstein, who was rightly impressed by good mechanical drawing, probably never considered the endless beautiful drawings submitted to patent offices for machines that cannot be made or which are wholly arbitrary (like the secret weapon drawing based on a vacuum cleaner in Graham Greene's Our Man in Havana). In another sense, picturing presupposes truth, rather than making it. For to describe the accident (a success notion), the picture has to be correct (true), so the whole project was circular.

Why should any kind of a matchup, no matter what it is, be truthmaking? Something else will always be required. No matter how successful our match-making, we need something to explain why that gives truth. And since there are various notions of truth in diverse subject areas, there could not be one kind of matching that is the same thing as truth; as a result, we'd have to explain why a certain particular kind of matchup would cause (or be) truth in one discourse and not in another.

It is easy to see why the idea of a representational matchup occurred to so many philosophers after the way of ideas became accepted. Elements of inner thought (particularly ideas of sense that are components of beliefs, judgments, etc.) were thought to represent outer reality and in true judgment, to replicate reality to the inner eye, and so, be true. So when I judge things to be as they are, my thinking or saying, matches, corresponds with, reality. How else could the "inner theater" display "outer reality"? It has to be like a good photograph or a good movie scene. But what is "good" if not "true"? That seemed much more plausible than saying what I think or say IS the reality because what I am immediately aware of are my own ideas, and I can think what no longer exists or does not exist yet. So a correspondence must be brought about by some "line-up" of the parts of thought with the parts of reality, roughly that the idea of the subject is related to the idea of the predicate in the same way that the denoted thing is related to the denoted predicate.

And so there comes an invitation, becoming more urgent with the re-invention of sentential logic and formal semantics, to provide the parts of both thoughts and facts, whose fit makes truth. Moreover in the background, even of modern analyses of statements, there are echoes of an older "two-name" theory of predication, which had various forms from Plato to Frege, and a particular medieval form in which variables are replaced systematically with proper names, and predicates are taken to be names of concepts (and by others to be names for extensions of particulars), so that a singular statement is true just in case what is named by the subject is a thing in the domain of the predicate, that is, "just in case what is named by the subject is one of the things named by the predicate," to use Ockham's version for the moment; on that pattern it is easy to map the rest of categorical statements.

For Ockham that is supposed to explain how a particular sentence is made true, and since judgments are made up of the natural signs attached to words, it is supposed to explain their truth as well. The judgment "Socrates is an animal" is true just in case Socrates is one of the animals. But "personal supposition," the medieval analogue of reference, but in reverse as "standing for," that holds between the natural sign "Socrates" and the person, is an intentional relation and without an explanation of that sort of thought state, there really is no explanation of truth at all. Similarly, Austin's account involving referential conventions and descriptive conventions simply smuggles in enough thought states to appear to explain something when it is the thought-content of statements that has to be explained; for that alone can be true. The same holds for another semantic account, "the statement s is true just in case the situation s expresses is a situation that is actual." How does a statement express a situation? It must be by our thinking. It is the true thinking that needs to be explained.

Consider an example. Suppose we attach a simple computer to the heating-air conditioning system in a house so that whenever the heat is off, the screen reads "the heat is off," and so forth, for each state of the system, including the temperature, humidity, and even when the system breaks down there is a diagnostic sentence. Now under Austin's analysis, as well as Chisholm's and Ockham's, every one of the sentences, the reports, will be true, and they will tell us what is going on. But, of course, there really is no saying at all; there is no judgment at all; there is just a simulation of saying and, thus, a simulation of truth. In brief, the reason the analyses I mentioned cannot explain truth is that all their conditions can be satisfied by a system that simulates truth, where there is no thought and thus no truth at all. Now of course, we interpret such readings as if they were assertions and, thus get knowledge from them. But that is only because we have built a system that is parasitic on our system for expressing statements in print. The problem remains: how to explain genuine truth that has not been confused with simulations and derivative and attributional forms of truth.

And who would think that the truth of the message could be explained by a matchup of the assembly of the signal (the uttered or written sentence) with the assembly of the reality? Even the best semantic accounts have a giant gap. And what about truths for which there is no independent external reality: say, of arithmetic, chess, poker, baseball and law? Are there such states of affairs that exist necessarily, for instance, that a regulation chess board has 64 squares? Suppose I think there are no Roman gods. How could there be a matchup of the assembly of the signal with the assembly of such a reality? Is the state of affairs, "there are no Roman gods," part of the necessarily existing furniture of the universe, rather than a content invented by our thinking? The ontology is grotesque: the marquee of infinite light bulbs some always "on" (the necessary states of affairs), some always "off" (the impossible states of affairs) and some "on" sometimes, and "off" others (the contingent states of affairs) -- the array forming state descriptions or possible worlds, with the actual world consisting of all lights that are "on": our world.

Now each of these bulbs is an abstract entity, a state of affairs; so the actual world consists of abstract entities! Where does matter come in? Can matter and contingent individuals be constituents of necessarily existing states of affairs? Next, there are apparently no truths that lack compliant realities, so if there are any truths of arithmetic there are necessarily existing, necessarily obtaining numerical states of affairs. So, too, for chess. But I say there are no such realities. We make the truth by composition, construction and invention of entia rationis. Such things have being only intentionally, and because of what is true of them, not vice versa. "Mathematics is the free invention of human thought." And with imaginary entities and abstract particulars (fictions, symphonies, operas), "being follows truth." There are no independent states of affairs for such truths. Similarly for "Quine never existed," had that been true! The above general formula for correspondence theories makes any theory that conforms to it ontologically false, and without explanatory force.

Correspondence theories that involve a matchup to make truth have not been held by many philosophers before Russell's and Wittgenstein's logical atomism. "Correspondence" has so long been offered as an interpretation of the phrase "adequatio mentis et rei" used by Thomas Aquinas, that it is often said that Aquinas, following Aristotle, held a correspondence theory of truth. Although Aquinas uses the word "correspondence" once in De Veritate, both his and Aristotle's understanding of truth (of judgment) was a sameness account: the knower becomes what is known.

If you had to match diverse things with distinct parts, say, judgments (made of ideas) and facts (made of atomic objects and universals), what relation other than a "same assembly," a congruence relation (as in picturing, mirroring, or modeling) could be truthmaking? Yet, paradoxically, why would parallel assembly or congruence be truthmaking, unless, of course the parts we use are the very parts of reality, and what is achieved is sameness? For, otherwise how could one assert or assent to what is so by asserting to or assenting to something that is only congruent with what is so -- and presumably, contingently congruent since the same thoughts are supposedly possible without the physical reality thought of. Of course, the latter supposition is absurd: you could not have thought Quine existed had Quine never existed at all.

What about John Austin's account of truth that he himself called a correspondence theory? Whether to count his four-place relation as a correspondence theory is disputable. Austin said "a statement is said to be true when the historic state of affairs to which it is correlated by the demonstrative conventions (the one to which it "refers") is of a type which the sentence used in making it is correlated by the descriptive conventions" [Austin, 1950, pp. 121-122]. My problems here are (i) the excess ontology of "states of affairs," especially of ones that do not obtain (but must exist somehow), (ii) the assumption that truth-valued judgment is linguistic, (iii) the implicit identification of truth as a property of statements rather than of thoughts (judgments); and (iv) the fact, described above, that all Austin's conditions can be met by simulations of statements.

Even if Austin's theory is considered a "same thing" theory, because the situation referred to is "of a type" with situations the statement correlates with descriptively, still, that won't escape the simulation. Besides, to extract the account of falsity is awkward: is a statement false either "when the historic state of affairs with which it is correlated by the demonstrative conventions is not of a type..., or when there is no state of affairs of the type to which it is correlated descriptively? I suspect we could neaten that up enough. But that would only expose the broken bones of a theory that assumes true judgments and sententially expressed, when most of our true and false judgments are not linguistically expressed at all. Some well known philosophers, like Donald Davidson and Marcia Cavell [1993] have committed themselves to claims that there is no truth or falsity without the use of language. But the many examples, ranging from reading words and music, to building, sculpting, painting and ordinary walking, as well as the invention of language, seem decisive to the contrary.

At first Austin's story looks better than a part-by-part replication; but then we see there is no reason why we should not go right through all the words to say: what he says is true just in case what he says is so! All the rest is just machinery that can be simulated by systems with no thought at all.

The same sorts of objections apply to R. Chisholm's [1977, p. 138] "correspondence" theory, which stripped of its technicalities [see Kirkham, 1995, p, 133 for a fuller statement] comes to saying: A particular sentence is true just in case the state of affairs that it expresses, obtains. That has the defects I mentioned: sentences are treated as underived truthbearers; it introduces a platonic ontology of states of affairs, and supposes that a sentence-token can express something definite apart from a judgment. But sentences don't: "The sun turns faces black." Without a context that sentence expresses nothing at all. Attempts to explain truth and falsity as semantic features of sentences are doomed; there is no truthmaking relation between words and the world. Sentences are vehicles for judgment. But not the only one. And above all, this so-called correspondence theory has no explanatory element in it. It passes by as something more than trivial because we have a background insight (needing a suitable psychology to explain it) that what I think is true when what I think is what is so. That's what needs to be explained. Moreover, Chisholm leaves out of account the plurality of our judgments that are not expressed verbally [see the discussions in Chapter 8, below]. If Chisholm or Austin had gotten the account of truth right, there would be no truth or falsity to my perception of the words I write or to your perception of written words.

Look again at the general formula Kirkham [1995, p. 132] proposed as a summary for both Chrisholm and Austin, and a general pattern for correspondence theories: "(t) {t is true if [(Ex)(tRx) & (x obtains)]}." Here "t" could be a judgment, a statement, a proposition, even a sentence. "R" could be "asserts that," "says that," "expresses that," but not simply, as Chisholm thinks, "means x" or expresses x in L" [Chisholm 1977, p. 138], because, outside formal languages, there is no such determinate relation. Chisholm thinks a sentence is true just in case what it means, obtains. That would make the simulations I described be true. But if no one uses a sentence to say anything, it is not true. The reason is that the "R" relation that connects truthbearer to compliant reality has to involve commitment, believing or saying. But skipping that, the whole scheme boils down to a claim that for a judgment to be true there has to exist a state of affairs that actually obtains -- an ontological necessary condition. I have already criticized that.

So to summarize. In one form, the match-up form, correspondence theories are just incoherent, empty promises of what no one can deliver; and in their more straightforward forms, whether Austin's or Chisholm's, they elaborately state what we already know without an explanatory advance, and are most unsatisfactory on the notion of falsity, all the while introducing ontologies, whether of infinite actual facts or states of affairs, or the like, that there is no reasonable basis for postulating at all.

2. Coherence.

Coherence has its uses in explaining truth. First, sharp notions of coherence involve bifurcation and a preference for affirmative over negative statements along with a background or base to begin with. Coherence cannot, however, in general, explain truth because it presupposes it. "Truth can't be coherence because we know the false can cohere" is a standard objection [e.g., C.J. Ducasse, 1944] that requires one to situate coherence against a base.

The originating illusion is formal thinking. Formal thinking fosters the idea that a certain kind of coherence assures truth, namely, the kind of coherence that makes every statement that is authorized by a certain form of thinking turn out to be fully satisfied by everything of which it is authorized to be said. But there can be that sort of coherence only if the objects, the designata of the generalized statements, are themselves formally generated by the sort of thinking used to manufacture the statements: e.g., numbers, geometrical objects, games, compositions, diatonic musical elements (notes, rests, meters, etc.). So while there are situations in which coherence, as logical consistency, can be sufficient as well as necessary for truth, given a certain base for starting the thinking, those are conditions that can be satisfied only in the case of formal and quasi-formal objects, including "mixed" objects like the elements of baseball in their "rules-aspect," but, of course, it is not coherence that makes the originating statements or base true. Moreover, whether a given assertion is authorized or not, e.g., a rule of chess, is not settled by coherence alone; after all, we could add a rule that a game is a draw after 158 moves.

Brand Blandsard [1941, p. 206], though he did not acknowledge the origination of the coherence idea in formal thinking, was a thorough idealist who thought that if the "physical world" is the product of thinking in certain ways, then consistency with the basic form of thought and non-contradiction would be what truth was. He did not seem to be sensitive to defining consistency. Nor did he examine whether the basic form of thought consisted of statements -- though I think he took it to be categories -- or whether more than one notion of consistency was being used.

As C.J. Ducasse [1944, p. 325] observed, "no coherent set of propositions can be comprehensive of all propositions," from which Ducasse concluded that for any coherent (consistent) group A there is an equally coherent but logically opposed group B. Ducasse was attempting to show that coherence as mere logical co-consistency cannot be sufficient for truth because groups A and B are both internally coherent and each is opposed to the other, so if one is true, the other must be false. Ducasse's assumed bifurcation might not be so welcome to an idealist today. And the role of a non-propositional basis, especially in sensation ordered by the refractive geometry of thought (e.g., Kantian categories, forms of intuition, etc.) was not properly recognized by the critics of idealism. For such a base for truth need not be a base of truths, but only what generates certain kinds of thinking. The new problem for coherence theory is to specify what coherence with a non-propositional, non-judgmental base is and how it can be violated. I think that can be done in an interesting way, but whether the limitation of supposing another notion of truth can be evaded, I am unsure. In any case, I think it unpersuasive to construct a vast refractive geometry of the mind without any explanation of the plurality of minds, the sameness of their refractive powers, the status of sensation -- all to avoid an identity account of knowing.

Coherence, as "being consistent with the basic structures, or basic postulates of thought" may have a place to explain the truth of some thought-systems with idealized objects. But "consistency," whether relativized to a base or not, cannot be defined without our using the notions of truth and non-contradiction; therefore, coherence cannot be a global account of truth. Nor can we be satisfied to use the anemic and contentless notion of "truth" we find in propositional logic for a general definition of consistency and a more robust "entailment" is a disputed and variously analyzed notion, not even part of propositional logic, and involves the notion of truth as well [see Ralph Walker, 1989, pp. 39, 144-45 and 192-93].

3. Pragmatism. Historical vs. empirical pregnancy.

Pragmatism, broadly, "what is true is what holds up in experience," has its uses. Pragmatism, that the truth of a belief (or a theory) is its working out in experience, or, in its epistemic form, its being approved by some (ideal?) community of appraisers, or fulfilling the expectations of some group, cannot be a global account just because "the final experience" in which approval is earned or expectation is fulfilled has to be the product of judgment that does not depend on some further working out, in order for the grounding of the truth to come to an end. It has to be true that the belief works out or is approved by its appraisers or expectation is fulfilled, and that cannot always depend on a further approval or authentication. Pragmatic truth has to wind down into truth by identity. Thus, it is a subsidiary notion.

There are kinds of "working out" that can be truth-making for limited domains, such as eschatological truth [Ross, 1988] and construction measurements; but only parasitically on its being true tout court, that expectation is fulfilled.

Pragmatic fulfillment is a truthmaker for certain kinds of historical statements where the truth conditions for what is said are not complete at the time for which the description is offered, except in broad outline that allows, even requires, a cultural fleshing-out (and where in fact the conditions of fulfillment are not contained in the then present facts); e.g., "He founded the Roman Empire." "Christ will come again in glory."

I do not mean only that historical statements have a predictive element, as if that were not found in ordinary physical object statements. An ordinary physical object statement that a certain pencil is green, is as predictive as a general hypothesis that emeralds are green. Most people don't notice the elaborate conditions of fulfillment, natural regularity, and the large class of observational hypotheticals that, if formulated, would have to turn out to be true in order for the pencil to be green [see Chapter 1, "Packed Virtuality"].

The range of things that is relevant to the truth or falsity of "the pencil is green" is by one test much narrower than the range of things relevant to the confirmation of "Julius Caesar was the first Roman emperor." For one thing what counts as "Rome" is itself a matter determined later on by history, as well as whether there was an empire, and later emperors. So it is no surprise that the conditions for historical truth are different from those for other kinds of statements.

Eschatological pragmatism is concerned with a certain kind of projective historical statement, and I do not mean by this just technical statements made by historians, but interpretative statements made by ordinary persons, that concern future human history, where the truth does not depend on what happens directly, but upon what happens mediated cognitively, that is, on the fulfillment of expectations, either for the predictors or for some group of successors or by some ideal observer.

If I believe humans will colonize space (or, Leonardo, that humans would make flying machines), the fact that the physical conditions eventually to be realized are unimaginable to me, or only schematically (as to Jules Verne) makes no difference if the later event would fulfill my expectations. My belief, in such a case, would have been true, otherwise not. We often cannot tell, as observers looking backward, whether expectation would have been fulfilled because there is a subjective element in expectation. Especially that is so with fulfillments that exceed all expectations. So the truth or falsity of earlier predictions and expectations is sometimes cognitively inaccessible to us. For many eschatological statements there are no determinate truth-conditions independent of variant expectations: "The Messiah will come and will establish the Kingdom of God here on Earth."

Note that such statements do involve a particular form of cognitive accessibility, namely, fulfillment of expectations, as a condition of truth and disappointment of expectational for falsity. H. Putnam developed a similar theory about science that E. Soza explains and criticizes [1993]. This is a different sense of "true," from, say, "It is true that we are humans," because crucial elements of what we mean by "being so" are different. For one thing, here "thinking what is so" is "expecting what is going to be experienced." It is clear that nothing can be true as fulfillments of cognitive expectations, unless there are things true in other ways. Thus, fulfilling cognitive states have to root, eventually, in realities. The final fulfillment of expectation cannot be a further cognitive state (except for trivialities: "In immortal life, our knowledge expands forever"). Pragmatic truth cannot be the basic kind of truth, but it can be a kind, not entirely reducible to sameness of thought and reality but rooted in it.

4. A transition.

"True believing is believing what is so" is analogous. What is rightness of understanding depends on what we are thinking about and the uses of thought. So also, is being so. A certain disposition of chess pieces is to be regarded as "check"; a certain hand of cards as a "straight"; a certain baseball situation as an "out"; those are distinct conditions for right understanding of situations. Far more subtle are the conditions for understanding human realities involving love, deception, support, abandonment, faith, and comprehension. Distinct, again, are the conditions of right understanding (right conception and judgment) for doing arithmetic, proving number-theoretic theorems, reading a musical score, weighing the force of circumstantial evidence, discovering analogies, and making deep metaphors.

Rightness of understanding is not one thing but a family of things, bearing marked resemblances in the predicates appropriate (true, false, erroneous, mistaken, correct, right, wrong, etc.). The different senses of "right" are not like "painted," applied to a painting and to the trim of a house; they are more like "collected," applied to debts and donations, shading off to be like "enjoyed" applied to classical music and a long life. "Is true" is not univocal; there are distinct families and, as I mentioned, distinct ways in which thought and reality can be brought to the same earned and inherited truth, and the most important case, inflationary truth (the truth of formal and quasi formal thinking).

The condition that explains a statement's (or other truth-bearer's) having a compliant reality varies with the area of discourse. The condition that is the "being so" also varies with our thinking, and, of course, with the analogy of being (analogia entis). No single analysis of "is true" will do for science, law, mathematics, logic, or, a fortiori, for theology, and certainly not for all parts of all; not, that is, if an explanatory objective is in view. All the same, analogous definitions, "truth is rightness of understanding perceptible to the understanding alone" (Anselm) or veritas est adequatio mentis et rei (Aquinas), "truth is thinking what is really so," whose meanings adjust to the context, can be useful. Such definitions are, as I said, linguistically as well as notionally analogous; their value is measured, in part, by their adaptability to what we see to be similar (and what some, unfortunately, find to be the same). Further, they focus what might otherwise appear to be unrelated things. There really is a range of right understanding, measurable by the mind alone, that is truth.

5. Sometimes the truth is what counts as so.

So, pragmatism is a viable subsidiary account of truth, for a limited domain: that the truth of judgment is the fulfillment of certain (historically conditioned) expectations by later cognitive states regarded as cognitively successful.

There is another analogue in scientific and ordinary judgments that are, implicitly of the form "that's good enough" or "that's what counts." So people measure a beam: "Exactly 12 feet," one reports. The other does not say "within how many millimeters?" because it does not apply as a test, where those intervals would apply in cylinder clearance for an engine. "Good enough" is "true" for practical purposes with adjustments all the time to the subjects and disciplines, neither denying that there could ideally in most cases be a more "refined" measure, nor denying that after one goes far enough, physical reality is indeterminate. So there are, practically, a lot of cases where "coming out close enough to expectations," governed by the practice, is what counts as true, and falling short (also practice-determined), is what counts as false. Piano tuning is a good case because perfect intervals are excluded from even-temperament, and beautiful tuning requires artistic adjustment of the middle octave and of the octaves, successively.

Finally, as mentioned, there are multiple families of judgments where "what is so" is the same as "what counts as so." So sometimes, the measure of truth is the measure of what is so, where that is determined by the requirements of the practice in settling what counts as so.

6. What about "eliminative," "redundancy" and "disquotational" analyses of truth?

Kirkham [1995] describes these views adequately I think, along with their difficulties. Basically, even if you could paraphrase away all uses of "is true" and "is false" you would not have torn out the problem. We are saying something with such predicates. "What?" and "with what justification?" remain. As I remarked earlier, if we simply dropped "true" and "false" and all their paraphrases as well, the problem would still remain to explain what is the truthmaking relationship (or sameness) between judgment and reality and how it can be lacking in false judgments.

The disquotational option is a triviality. Even skipping purported and contested counterexamples for now, "(p) is true" cannot always be reduced to or replaced by simply asserting P. Isn't "semantic ascent" a way of talking about the world by talking about what we say about the world, as Quine suggested [1970, pp. 11-12], see Kirkham [1995, p. 319]? And there are cases where object-language expansion would not do at all. Secondly, in law and religion and elsewhere, we use "is true," etc., to elicit or express commitment not only to what is said but to the words in which it is said, and, as mentioned, to affirm the veracity of what is said, sometimes without commitment to facticity. Thirdly, the correlate "(p) is false" cannot decompose into not-P because if Quine never existed there would be no not-Q. And that holds for judgments involving a proper name or a real kind word in general.

Quine was right to say (of sentences), that we need the "is true" and "is false" as predicates to make certain claims about infinite lots of them, and even finite lots of them. Another point Kirkham [1995, p. 320] makes about typical uses of "it's true that" is that the pragmatic force includes, frequently, an acknowledgement that the "proposition has already been uttered"; that is certainly so when the "it is true that" is part of a concessive introduction to an ensuing disagreement: "what my colleague says is true, but...." To boil this all down, there is something predicated of a judgment when we say it is true and that is that what is thought is the same as what is so, adjusted for discourse and reality. No semantic device for eliminating the need to use the word, will eliminate the need to provide the explanations. Just the fact that thinking and saying can diverge from what is so requires an explanation of what a thought's being so consists in.

7. Local roles.

I will explain in Chapter 5 that there is a basic knowing and believing in which the constituents of thought are real things and real conditions present to intelligent awareness, made thus present by abstraction, transforming animal awareness. But that does not mean that all judgment has to be of that sort. There are families of representational knowing, parasitic for success upon the existence and priority of the basic judgments described. A diverse multitude of experimental judgments involve representational knowing: for instance, that we are presented with an appearance which we know, for one reason or another, is not "the real appearance" of what we are to judge, and yet, from that appearance we can make, and check, reliable judgments. How that happens is varied. From an overhead photograph of a crowd, with known dimensions for the area, specialists can make accurate judgments of the number of persons. From photographs of the night sky, where none of the apparent positions are the "true" positions of the stars (including planets, galaxies, gas clouds, etc.), experts can construct the "true" positions which would look quite wrong to star gazers.

Sonograms use a different process, producing an image by sending sound signals through external flesh and muscles to internal organs. The detection of a kidney stone or a blockage is not really by seeing the organ, but by seeing an effect (the image), features of which reliably correlate with features of the organ to a trained reader. That is established by pathology. In general, knowing by probing, whether the sense be visual, auditory, tactile, olfactory or with instruments by touch, in which an effect is observed to reveal features of its cause is a perfectly intelligible means of knowing through representations. Yet all such knowing is subsidiary to knowing by direct presentation [to be partly explained in Chapter 5], without which there would be no such thing as knowing causes through their effects.

Similarly, there are neighborhoods of discourse in which "working out in experience," e.g., yielding close to the predicted values in measurement, is not only sufficient for warranted assertion, it is sufficient for truth. What we say when talking about engines, pumps, pipes, heating systems, and the like, involves notions of "true" where "work out" or "count" is what we mean. Now of course that sort of truth presupposes a more general notion, close to "is really so," but for practical purposes the more immediate "yielding the required measurements" not only will do, but is what counts. The effectiveness of such practices depends on prior notions where the realities judged are the elements of the judgments.

Again, we sometimes employ contextual notions of coherence for what counts. Suppose we are checking testimony from an expert's deposition: if we grant at the outset that certain claims are so, we can scan the rest for coherence, using a consistency test first, then thickening the test, to include "does the same explanation apply throughout" and adding additional subject-specific tests for coherence, e.g., "Does that fit the building practice," etc. When all are satisfied we can reasonably conclude that the rest is true, too. That's just a rough and ready example of situations in which, given grounded initial convictions, we can proceed to others by tests for coherence or tests of satisfied expectation (which are subject-specific). But even used that way, coherence and pragmatic effectiveness is at most another meaning of "true," subsidiary and ancillary.

8. Emotivism and other add-on theories of truth.

Emotivism can be dismissed, as well as redundancy accounts that implicitly deny that "is true" is a predicate. "Is true" is not a mere intensifier applied to a claim, nor an implicit "I know that" or "I am certain that" operator, nor any of the other "add on," explanations, though it can be used in those ways. These are among the incidental functions of "is true," not a general analysis of it.

9. Parasitic locality.

Notions of truth, e.g., "working out in experience" or "what counts," that cannot be global notions of truth, can and do function in various practices, like, engineering, medicine,and building. That is another reason to consider the notion "what is really so" to be analogous and not to require that "what is really so" has to be independent in its existence or its features from what we, or some of us, think, or has to involve some reality that is complete apart from human thinking. Now that could not be true in general, but it has to be for entia rationis. Similarly, that there are 64 squares on a regulation chessboard is really so, though there is no such reality apart from human thinking, and in this case, agreement. There is objectivity without independent reality; the same is true for arithmetic, set theory, logic, checkers, and even the game, baseball.

Similarly, mathematics is as much the product of human thinking, inventing the conditions of authorized assertion and satisfying internal requirements, as are musical composition, poetry, and science fiction. Yet we still say, "what is true to think and say about whether every number that has a cube-root also has a square-root is what is really so." The reason is that the right answer is independent of individual opinion and the same for all who do the thinking properly. So, "what is really" so has to apply analogously, and, thus, so does "what is true to think and say."

V. What Does the Failure of All Those Theories of Truth Tell Us?

Some philosophers, like Russell, Wittgenstein, and Frege, took "what is the general truthmaker for our thinking?" to involve a thought/reality match, instead of a thought/reality identification. Also they had an unanalyzed notion of causation in mind, one for which they could provide no analysis. How can reality make a representation of it true? Further, many speak of sentences as true or false, of the truth conditions for sentences, and as if truth conditions were the same as the meaning. They ignore the endless overflow of truth conditions that are determinate only by referential practices. And many talk as if sentences or utterances were somehow capable of truth or falsity apart from the judgments expressed, so that a computer, programmed to make sentences, could produce an endless supply of both true and false ones. Some thought there might be a semantic feature of true sentences to distinguish them from the false, and looked for some relationship between words and the world that would account for truth. Those attempts met frustration and defeat.

With the failure of matchup theories obvious to everyone, and people still convinced that thoughts are made up of representations of things (whether that be images, linguistic elements, impressions of sense, or whatever, as long as the components are not the things themselves and their conditions), the strategies for explaining truth became severely limited. If thoughts are made up of thought parts that are never real things or the conditions of things, then either truth-making is a relation among thoughts themselves, or reality has to be somehow the product of thinking or truth has to be some kind of confirmation of imagination by later experience. Worse, if you are a materialist then you want the story to go full circle and in its second stage to explain how thought is nothing more than the matter which the earlier stage says is made by thought. This is tailbiting.

So, we have to rethink the matter. We have to reconsider that (i) in thinking we transform the physical, by making it present in thought and subject of judgment; and that (ii) the basic, but not the only, constituents of thought are real things and the conditions and the changes of things. And, secondly we have to abandon the futile conviction that truth and falsity are semantic features of elements of language. And, also, the idea that truth is ideally warranted assertability because the analysis will not pass its own test.