We have no doubt that other people and maybe higher animals have mental qualities.

Since physics is content with knowledge of structure, belief in the minds of others requires some postulate not required in physics. What further postulate is needed?

We must appeal to “analogy”. Analogous behavior must have analogous causes.

This isn’t just the claim of escaping solipsism. This goes further, so that we can know the thougts and feelings of others. Of course, such knowledge has its objections (dreams, robots, calculators, grammophones). The illusion of a mental life is able to be produced.

But how do you know it’s an illusion? How do you know the grammophone doesn’t think?

There is a difference in the causal laws of observable behavior. Computers cannot account for information their programmers didn’t account for. Also, Russell says, humans peculiarly are able to change their response to a given stimulus.

But this difference in behavior doesn’t suffice to prove the existence of other minds. A person’s behavior can probably be accounted for with only causal laws, and materialism is probably impossible to refute from external observation alone. Some extraphysical inference is required to conclude the existence of other minds.

Russell wishes to discuss the possibility of a postulate that establishes a rational connection between a belief about others’ minds and data.

This abstract schema seems to be: We know from self-observation “A causes B”, where A is a “thought” and B is a physical occurrence; an observation of B in cases where we cannot observe an A leads us to infer an unobserved A. Increased complexity of data and increased certainty of the causal law leads to an increased confidence in the inference.

Sure, such inductive arguments are, in their own way, suspect.

A simple cause (say, “thirst”) accounting for complex observed behavior increases the probability of the inference by also decreasing the probabilities of other inferences. In ideally favorable circumstances, the argument would formally go as:

I know from self-observation that A causes B, and only A causes B. I could be wrong that the causal law holds outside my experience. The postulate assumes probability rather than certainty in such cases. The postulate can thus be reworded:

If, whenever we can observe whether A and B are present or absent, we find that every case of B has an A as a causal antecedent, then it is probable that most Bs have As as causal antecedents, even in cases where observation does not enable us to know whether A is present or not.

This postulate justifies the inference to other minds.

2. The Qualitative Properties of Time, p. 19-24.

To establish that time is accounted for by physics, we must list those properties peculiar to time’s structure in everyday experience.

We begin by distinguishing quantitative and qualitative properties of time. Measuring time with clocks uses time’s quantitative or metrical properties. This includes determining equivalent time lengths and also determining simultaneity of spatially separated events. The theory of the metrical properties of time has been developed in great detail, particularly by Einstein and his antecedents. For discussion of this, refer to Reichenbach 1921, 1928, 1969.

Here Reichenbach focuses on the qualitative or topological properties of time. They are more fundamental in that they don’t depend on methods of measurement in any way. These properties are what give time its specific nature different from space and account for our emotive attitude toward time. So, a series of statements comprising the most obvious properties of time:

1) Time goes from the past to the future.

This is time flow, or becoming. We susually speak of time flow as changes in some objective entity we perceive and continual inexorable escaping. But the flow of time could remain connected with the structure of human consciousness; though derived from an objective root, a common product of an objective factor and a subjective factor. This question will be answered by a more detailed consideration of the physical root of time.

2) The present, which divides the past from the future, is now.

In one sense, this is trivial. In another, it is enigmatic, for what is the meaning of “now”? Is it a time point distinguished from other time points, or does it merely reflect our subjective approach to time, analogous to the space point “here”? We can choose “here”; can we choose “now”? There are restriction, leading to more statements.

3) The past never comes back.

This is closely connected with time flow, with (1). It expresses the direction of time flow, and precludes closed time-like loops, making time a one-dimensional or linear continuum.

The next three statements concern specific differences between the past and the future.

4) We cannot change the past, but we can change the future.

Of course, our ability to change the future is limited – we cannot change most things. More precisely, (4) says: There are some future happenings which we can control, but there are no events of the past we can change.

We might be tempted to formulate the knowability difference, but it is obviously false that the past is known and the future unknown. The difference, rather, is with respect to the way in which we acquire such knowledge.

Knowledge of the past is based on records, but records of the future seem an absurd concept. Isolated data from which records are made aren’t enough to predict the future.

Future forecast requires comprehensive information covering the total occurrence.

So we only have registering instruments for past events. They are controlled by partial effects of comprehensive occurrences which allow us to infer the larger occurrence. Registering instruments for the future seem absurd.

Isolated indications can allow prediction of the future in exceptional situations, when observations imply that a comprehensive process is going on, like a foundation crumbling. But the predictions in such cases are general – they cannot yield specifics like where exactly all the pieces of the house will fall once the foundation collapses. Predictions require a knowledge of the total cause; postdictions can be based on partial effects, records. We formulate this as:

5) We can have records of the past, but not of the future.

We can sum up (4) and (5) with:

6) The past is determined; the future is undetermined.

What is the meaning of “determined” and “undetermined”? We could say that it means the past consists of established facts which we cannot change and can be recorded. But this leaves the question of how far (6) goes beyond (4) and (5).

This will have to be examined, but we’ve left other questions unanswered, too. Many of the terms we’ve used require an examination of their meanings.

Therefore, this analysis of time is a clarification of meanings. It is a problem of explications, where we replace the explicandum with the more precise explicans that justifies the use of the statement in the context of human behavior.

An explication can never be strictly proven, because the explicandum is vague. It can only be adequate, i.e., replacement doesn’t change truth values in most sentences. If so, then the explication is a proposal to use the precise term in place of the vague one. This helps us arrive at precise meanings and formulate relations to the concept that otherwise wouldn’t have been known.

Explication of the concept of time, with its derivatives like time flow, time direction, &c., is of particular significance because of the major role time plays in science and everyday experiences. Sch an explication will require studying some fundamental physics and will clarify the foundations of physics and philosophy. That this is important to philosophy is obvious, because, as long as time isn’t understood, the philosopher cannot claim to give an account of physical reality or human knowledge of it.

Body events influence mental events. Characteristic example: eat a Hershey bar, taste chocolate.

Evidence shows mental events depend on brain events. It isn’t clear how, but it is clear that it does.

We do know in some cases how the mind and brain affect each other. Examples are given. While unclear on all the details, we know these relationships are very complex. This, so far, is all science, not philosophy.

There is, though, a philosophical question: “Is your mind something different from your brain, though connected to it, or is it your brain?” Or, rather, are the so-called mental events part of the physical processes of the brain, or are they in addition to these processes?

Example: chocolate bar. Physical processes eventually lead to you tasting the chocolate. “What is that?” Is the tasting just a physical process or event in the brain, or does it have to be something completely different?

If a scientist removed your skull and looked at the brain while you were eating chocolate, he would see only brains, and would only detect chemical processes. Would he find the taste of chocolate?

Even when observing the brain, a scientist can’t find your experience of tasting the chocolate. Experiences lie inside the mind with a kind of insideness that is different from the way that your brain is inside your head. They can’t cut into your mind.

It’s not just that the scientist can’t see flavor – even if he licked your brain, it wouldn’t taste like chocolate. Even if it did, he would only find his experience of chocolate, not your own experience of chocolate.

If experiences are so different from brain processes, then it doesn’t seem they can be physical states of your brain. There hse to be more to you than your body and nervous system.

One conclusion is that there is a soul attached to your body such that the body and soul can interact somehow. If so, then you’re made of two different things: a complex physical organism and a purely mental soul. This position is called dualism.

Many hold this to be unscientific. Everything else is physical; why shouldn’t we be? Couldn’t the complex physical system that is our bodies be sufficient to give rise to mental life? Could a mere philosophical argument show otherwise? How can philosophy tell us what something’s made of?

This view, that mental states are just physical brain states, is called materialism, or physicalism. Materialists don’t have a specific theory of mental processes, but only believe that mental states are physical brain states, and there’s no philosophical reason to think otherwise.

Materialists hold that we’ll discover the real nature of mental processes in the future, just like we did with water. We discovered that water was really hydrogen and oxygen, very different from water individually.

So, mental events being physical processes wouldn’t be any stranger than lots of other things we’ve discovered. It’s only a matter of time, say the physicalists.

A dualist will reply that the materialist is missing the major difference. The other things are clearly out there in the physical world. “It is an essential feature of this kind of analysis that we are not giving a chemical breakdown of the way water feels, looks, and tastes…. The physical… analysis of water leaves them aside.”

To discover taste is a brain process we would have to analyze something mental in terms of physical parts. And there is no way that a large number of physical processes can be the parts of a taste sensation. A physical whole can be analyze3d into physical parts, but not a mental whole.

There is another view that differs from both dualism and materialism. Your mental life could go on in your brain, but all those experiences, feelings, &c. are not physical brain processes, i.e., the brain is not just a physical object. It is itself both physical and mental.

This is called the dual-aspect theory. Experiences and other brain/mental processes have two aspects: a physical one and a mental one. Scientists can observe the physical aspect, but not the mental one – only you will observe that. Your brain has an unreachable inside. Certain brain processes will feel a certain way to you, and to you alone.

Another way to say this is that you are just a body, but your body isn’t entirely physical. It has two aspects, the physical, public one and the mental, private one.

Physicalists believe only the observable, studyable things are existents. But they have to account for mental events like feelings, desires, and the like.

Some argue that the “mental nature of your mental states consists in their relations to things that cause them and things they cause.” E.g., what makes something a pain is that it is the kind of state of your brain that is usually caused by injury, and that usually causes you to yell and hop and avoid the thing that caused an injury. And that could be purely physical.

But that doesn’t seem sufficient. Pains also feel a certain way, which is different from their relations to causes and effects. Nagel believes inner conscious experiences cannot be adequately analyzed in terms of causal relations to physical events.

There seem to be two very different things going on: public physical events and private mental events. This is probably true for animals, too.

We won’t have a good idea of the world until we can explain how consciousness emerges from a complex biological organism. If consciousness could be identified with a physical state, a unified physical theory of body and mind would be forthcoming. But a purely physical theory of mind is unlikely. There may be more to the world than can be understood by physical science.

1. The Emotive Significance of Time p 1-17.

Reichenbach sketches many of the common questions associated with the nature of time. He warns that these questions reveal the highly emotional content associated with the experience of time. He stresses that we should follow a more scientific approach, looking for answers to questions which have clear meanings. Only then can these emotional questions be answered.

He shows how the history of philosophy is filled with examples of this process of clarifications of meanings.

Inquiry into the nature of time has a similar history, puzzling the ancients and others for 2000 years, then finding an answer indirectly via the problem of causality. Examining the older theories helps reveal the emotional reactions and reveals the logic puzzles we encounter in our experience of time.

We cannot control the flow of time; it passes away inexorably. Our actions can only be directed at future events, but future events are uncertain. We can know the past, but we cannot change it. Death is the inescapable result of the flow of time. Fear of death becomes thus fear of time.

Problems which are claimed to be unable to be solved by means of logic are dissatisfied emotions projected into logic. The fear of death has greatly influenced the logical analysis of time. So-called paradoxes are defense mechanisms against physical laws that have aroused deeply rooted emotional antagonisms.

Religious philosophers assert the existence of a higher, non-temporal reality out of a desire to survive death. Parmenides and Plato also developed such theories.

The paradoxes of Becoming are an attempt to appease the desire to escape the flow of time and to allay the fear of death. Parmenides argues something would have to grow from nothing. Zeno’s paradoxes are supposed to demonstrate the impossibility of motion, and therefore the timelessness of Being.

Reichenbach answers Zeno’s two most famous paradoxes. These answers require a theory of infinity and limits achieved in the 19th century. So Zeno’s paradoxes are historically important in pointing out the problem.

Of psychological intereset, though, is that these paradoxes weren’t discovered by logical analysis of infinity or continuum, but because the Eleatics wanted to prove the unreality of time. He could have argued that since arrows do fly, or that because Achilles can overtake a tortoise, then something must be wrong with the logical process. Bt he wanted to escape time and death.

Where Parmenides embodies a negative emotional attitude toward the flow of time, a positive attitude, in which change and Becoming are an inexhaustible source of new challenges, was embodied by Heraclitus.

“All things are in flux”, Heraclitus says, and that’s okay. A physics of things doesn’t require a denial of time-flow, agreeing with common sense.

Of course, Heraclitus’ attempts to equivocate oppsites doesn’t really help argue for his position or add much to the logical analysis of time. An alleged logic of opposites cannot solve the problem of time and Becoming.

Heraclitus is naive, but many regard his approach as the turly philosophical one. Darkness of language has too often been the guise of triviality, falsehood, and nonsense. A clarification of the meaning of time and Becoming can be expected only if questions raised by common sense are answered with the help of the scientific method. The analysis of time had to be connected with the analysis of science in order to become accessible to logical clarification.

The study of time is a problem of physics. Emotive reaction cannot tell us what time is. Subjective expreience cannot tell us enough information about the time order connecting physical events. Reichenbach gives evidence of this. What we wish to know is whether our emotional reaction is justified, whether there is an objective time flow from future to present to past. For answers, we must turn to physics if we want anything but a psychological understanding of time.

If the past is distinguished from the future as the unchangeable from the unknowable, does this mean the future is still chageable? Daily experience implies yes.

Scientists may question the changeability of the future. Unknowable is not necessarily indeterminate; perhaps the future is as determined as the past, making the difference between past and future simply one of knowability. Time asymmetry would then be a matter of knowledge and ignorance; time itself would be symmetrical. This is suggested by science because science has accepted the universal validity of causality.

We see causal laws at work in the past and we also see them confirmed by future facts we’ve correctly predicted and later became reality. The future is not entirely unknowable. What led some philosophers to question the reality of time is the fact that some undesirable future events (i.e., death) are predictable. Why not assume the future is determinate?

Causal determinism was developed in antiquity. There appeared fatalism, in which ends are determined but ways lare left open, and the position of Democritus, where ends are irrelevant. But genuine causal determinism arose with modern science.

The physics of Galileo and Newton was exceptionally praxic and precise.

It is no wonder that determinism became an increasingly influential doctrine. Newton’s physics seemed to govern everything; Laplace didn’t hesitate to assume this. Thus he concluded that the future was just as determined as the past. He formulated the complete symmetry between past and future in his famous remark about a logically possible superman:

We must consider the present state of the universe as the effect of its former state and as the cause of the state which will follow it. An intelligence which for a given moment knoew all the forces controlling nature, and in addition, the relative situations of all the entities of which nature is composed – if iti were great enough to carry out the mathematical analysis of these data – would hold, in the sme formula, th emotions of the larges bodies of the universe and those of the lightest atom: nothing would be uncertain for this intelligence and the future as well as the past would be present to its eyes.¹

This is the classical formulation of determinism and would, were it true, spell the breakdown of a realistic interpretation of time flow. For causal determinism, as for Parmenides, there is no Becoming.

Causal determinism has become even more explicit and precise with the advent of Einstein’s and Minkowski’s four-dimensional spacetime manifold. Minkowski spacetime is a four-dimensional Parmenidean Being, in which nothing happens. Time flow is an illusion, Becoming is an illusion; we experience time this way, but it doesn’t correspond to reality.

Reichenbach’s comparison of the deterministinc conception of time flow to the happenings in a movie theater.

The paradox of determinism and planned action is a genuine one. Reichenbach promises a more detailed analysis, but points out now that if time has no direction, planned action appears incomprehensible. We can plan for the future, but planning for the past is senseless. Planning presupposes time flow, or adleast an acceptance of time flow. Can scientific analysis support this conviction?

The criticism of time flow developed by the ancients can be overcome by an improved theory of infinity. The objections from a deterministic physics are much more difficult to refute. A detailed analysis will follow, but first Reichenbach examines a modern revival of Parmenides in full knowledge of determinsitic physics, namely, the theory of time given by Kant.

The historical function of Kant’s philosophy is to formulate in logical terms what earlier speculative philosophers had attempted to say in mystical similes and aphorisms. Kant’s synthetic a priori is an attempt to characterize the contribution of human reason to knowledge. He thus revives in logical form the theory that reason governs the physical world – a rationalistic theory which had been, in Plato for instance, the creed of the mystic. A general treatment can be seen in Reichenbach 1951; we’ll restrict ourselves to Kant’s theory of time.

The Parmenidean distinction reappears in Kant’s philosophy as the distinction between timeless things-in-themselves and time as the subjective form of things-of-appearance. Kant probably would have objected to linking “subjective time” with “illusory time”, but it can be seen that “subjective time” is a logically more precise term for “illusory time”.

Kant presents long arguments in favor of his conceptions. Time flow, change, is not a property of Being, Parmenides said, and Kant agrees, adding that time is merely the form in which we experience Being.

How does Kant come to this conception? Like Parmenides, Kant wanted to escape time flow and death, which is obvious from his insistence on immortality. But he also wanted to save freedom of will, planned action, so that he could make man responsible for his actions – how can one establish morality if human action isn’t free?

Kant was faced with a highly successful deterministic physics. He believed he found a way out of determinism by teaching the ideality of time. If determinism as formulated in Laplace’s simile of a superman holds, then time is subjective, there wouldn’t be the possibility of time flow. Kant’s solution is to trade th ereality of time flow for freedom of the will.

Kant’s treatment of causality is important here. He followed Hume in saying a sequence of perceptions does not establish a time order of physical events; we must know causal order.

But causal order is just as subjective as time – it doesn’t express a property of things-in-themselves. So physics can only tell us about how the world appears to us, not how it is independent of human experience. And determinism refers only to the world of appearance, i.e., the world subject to causal order. The world-in-itself is free from the rule of causal laws.

Strange – time flow subjectivism leads to causality subjectivism. To gain freedom and morality, Kant sacrifices physics as a science of objective things. He goes beyond Parmenides in exempting Being not only from time flow but also from causal determination.

(This seems to me to be th elogical next step of any kind of time flow subjectivism, from Parmenides through modern quasi-physical BS authors *hint hint Greene*. -Chase)

It’s hard to see how this re-establishes freedom. Action is intended to change temporal happenings, but they are subject to causality, and therefore sceintific determinism. What kind of influence can we hope to have over them? Supposing an atemporal world doesn’t help, because we have no concern with affecting such things. We are only interested in changing things that are subject to causality and those that are perceived in our experience. We use causal laws to do so, knowing that plans would be pointless if there were no such laws. What is a philosophy good for, if it evades answers to questions about what men can do, by telling us that there is another realm of Being which we certainly cannot control? It comes down to an escapism. It does not solve the paradox of freedom and determinism, clarify the experience of time flow, account for the distinction between past and future, or between the unchangeable and the realm of what we hope can be changed.

Common sense strongly supports a distinction of past and future and an objective interpretation of time as a process of Becoming. But this cannot be accepted without philosophical criticism. Reliance on intuition often turns out to be misleading.

Escaping into common sense is just as much a pitfall as escaping into metaphysical speculation.

This applies to Bergson also, the modern ancestor of Heraclitus. Bergson also reacts to time flow with a positive emotional attitude. To Bergson, as to Heraclitus, the idea of Becoming offers emotional reward.

Such an attitude may be a good help in starting a logical investigation, but it cannot replace it. Appealing to intuition, to the “immediate data of consciousness”, does not say very much to someone who wants to know whether he can trust his intuitions. The meaning of Becoming can be clarified only by logical analysis. An act of vision is no substitute.

There is no other way to solve the problem of time than the way through physics. The physicist must either have discovered the objectivity of time flow and Becoming, or else have been able to ignore time in his contstruction of reality and describe the world without the help of time. Previous theories (Parmenides, Kant, Heraclitus, Bergson) are insufficiently grounded. If there is a solution to the philosophical problem of time, it is written down in the equations of mathematical physics.

Perhaps it’s more accurate to say that it is in between the lines of the physicist’s writings. For instance, there’s no physical equation for determinism. Philosophical analysis is concerned with statements about the equations of physics. The philosophical analysis of physics is given in the metalanguage of physics.

Modern quantum abandons classical determinism. It will be important to investigate the implications of this turn to indeterminism, but there are other things to look at first. Thermodynamics is explicitly concerned with the problem of time flow. Certain physical equations that are very fundamental express the specific nature of time as distinct from space. The quantum contribution can only be understood after a study of the thermodynamical approach to the problem. It will turn out that physics can account for time flow and Becoming, that common sense is right, that we can change the future, but the proof requires the scientific method, even to understand the meanings of the terms “time” and “Becoming”.

¹Laplace, Pierre S., 1814. Essai philosophique sur les probabilités, “De la probabilité”.

What properties must a language have in order to be learnable?

Donaldson argues against a purely a priori account of language learning. He takes the position that it must be possible to give a constructive account of the meaning of the sentences in the language, i.e., a theory of meaning for the language. A theory of meaning that conflicts with condition, he argues, isn’t a theory of a natural language. If it ignores the condition, then it’s missing something. He also shows that several such theories of meaning do so conflict or ignore this condition for being learnable.

I. p. 3-7.

Davidson gives a brief caricature of the building-block theory of language learning espoused by the empiricist epistemology. He then shows that it is questionable at best and that the underlying epistemology has been rejected.

Davidson cites Strawson as holding a language learnability theory stemming from the building-block theory. Strawson attacked Quine’s view that “the whole category of singular terms is theoretically superfluous”. Strawson says that being able to paraphrase singular terms doesn’t imply the theoretical possibility that we could speak a language without them. Then he tries to establish the theoretical impossibility of this.

Davidson isn’t arguing against two of Strawson’t theses: that the eliminability of singular terms doesn’t follow from paraphrasability, and that eliminability is impossible.

Davidson instead wishes to contest Strawson’s argument against Quine’s eliminability thesis. It has two essential claims:

1) for any predicate to be understood, some predicates must be  learned ostensively or by “direct confrontation”;

2) for such learning to take place, the ostensive learning situation must be “articulated in the language” by a demonstrative element which picks out or identifies entities of the sort to which the predicate applies.

Quine has contested that (1) and (2) doesn’t establish the necessity of singular terms. But Davidson is interested in the implication of (1) and (2) that substantive questions about learning language can only be decided a priori&lt.

Surely it is empirical whether one gains abilities as a result of experience. But (1) and (2) claim it is a purely “logical” matter that everyone who has learned a language has done so in a certain way. For Strawson, ostensive learning has a predicate = learning by direct confrontation. There are two ways to take this where sch a process is intended to be more  special than learning the meaning of a predicate through hearing it used in sentences. One is to require the teacher to intend to bring the object to the learner’s attention, but this hardly seems necessary. The other is that direct confrontation and ostension require the presence of an appropriate object, while the correct use of a demonstrative singular terms does not. But it isn’t a priori, and probably not even at all, true that someone couldn’t learn their first language in a skillfully faked environment. Even Strawson admits that is too limiting.

Davidson thinks Strawson’s defense of (2) reveals the confusion underlying this argument. There is no reason why ostensive learning of predicates must be “articulated in the language” in one way rather than in another. There is no reason why reference to particulars must be by way of demonstratives. Examples from Strawson’s article reveal that Strawson thinks that Quine’s eliminability thesis is false because we would make a major conceptual change – alter the meaning of all retained sentences – if we cut them off from their present relations with sentences containing singular terms (or demonstratives, or proper names, &c.). Davidson suggests that Strawson erroneously or illegitimately draws conclusions concerning the mechanism and sequence of language acquisition from arguments for the conceptual interdependence of various basic idioms.

Quine agrees to a point with conceptual interdependence. If such claims of interdependence are tenable, then we couldn’t learn a language with predicates like ours and no singular terms, because there theoretically can be no such language. The lesson: insofar as we take the “organic” character of language seriously, we cannot accurately describe the first steps towards its conquest as learning part of the language; rather it is a matter of partly learning.

II. p 7-15.

We can describe what it means to know a language in a logical fashion. One natural condition to impose is that we must be able to define a predicate of expressions based on their formal properties that selects the class of meaningful expressions, i.e., a grammar for the language. Another condition is to be able to specify based on formal considerations what every sentence means, with semantical structure induced onto the sentences by the speaker’s actions and dispositions. This is similar to Tarski’s structure for determining truth-value. The theory doesn’t have to have truth-value for its “meaning” definition, but that is sufficient to work.

We define: a semantical primitive is an expression for which the rules that provide the meaning of sentences in which it doesn’t appear we are insufficient for providing the meaning of ones in which it does appear. Then the condition becomes that a learnable language has a finite number of semantical primitives. This is rough, but sufficient to point out flaws in other theories of meaning.

First example. Quotation marks. We should be confused by quotation marks. They seem to follow a rule, but are problematic when we try to incorporate it into a theory of meaning.

Quine, Tarski, and Church all have misgivings about quotation marks. We are, I suppose, tempted to think of them as a functional expression because their use on an expression denotes that expression. But then we must treat the denoted parts as singular terms or variables. This is problematic, and we must give up the idea that quotations are “syntactically composite expressions”, made up of the quotation marks and the quoted expression.

Tarski and Quine want to treat the quotation-mark names as simples, the parts of which are independently meaningless. But….

There would then be no theoretical justification for the classification – it is only an accident that quotations are “spelled” the same way. There would be no significance in the fact that a quotation names its “interior”. And there are infinitely many quotations, so a language treated this way is unlearnable.

Intuitively untrue. Quotations <i>do</i> seem to have rules, implying that quotations have significant structure, so it’s hard to deny that there must be a semantical theory that explains it. And neither Tarski nor Quine seem to want to deny the possibility. Tarski doesn’t endores his above statement, and elsewhere seems to consider the structure of quotations.

Quine argued that the lack of logical structure of quotations could be dispelled by changing notation. As such, sentences with quotations could take their meaning as equivalent to sentences so changed. This removes the infinitude of semantical primitives, and while not making quotations exactly “logical”, then at least directly linked with the logical.

Second example. Schaffler on Indirect Discourse. Schaffler sidestepped an argument between Carnap and Church using his inscriptional approach to indirect discourse. He suggests we analyze “Tonkin said that snow is white” as “Tonkin spoke a that-snow-is-white utterance”. The expression “that-snow-is-white” is to be treated as a unitary predicate of utterances (or inscriptions). Like quotations, the syntax is fairly clear. But it isn’t clear how the meanings of these predicates depends on their structure. So, without new theory, each new predicate must be a semantical primitive, failing the learnability condition.

It could be that more theory will reveal more structure, but then Scheffler would have to abandon the claim of having shown “the <i>logical form and ontological character”</i> of sentences with indirect discourse. Quine’s trick with quotations won’t work here.

Third example. Quine on Belief Sentences. We could just declare the problematic expressions as meaningless parts of meaningful expressions. “But semantics without ontology is not very interesting, and a language like our own for which no better could be done would be a paradigm of unlearnability.”

Concerning belief  sentences, Quine almost goes this far. Near the end of a search  for a theory of belief sentences he remarks that once we give up trying to quantify over things believed we can instead treat “believes that” as an operator, which produces a composite absolute general term including the sentence as a constituent.

In one respect, this goes beyond Scheffler, making main verbs (“believes”) inaccessible to logical analysis. Constituents and operators are syntax, not semantics. If anything aside from syntax is common to belief sentences, Quine doesn’t indicate it. This also makes it unlearnable. There could be further theory, Quine says, like for quotations. But it would almost certainly have to add to the ontology.

Fourth example. Church on the Logic of Sense and Denotation. In the last two examples, frugality of ontology led to problems with the desirable minimum of articulation of meaning. But prodigality in ontology won’t fix it all, either.

Frege proposes that verbs like “believes” do double duty. 1)First, they create a context for the following expressions to refer to their normal meaning. 2) SEcond, they map persons and propositions onto truth values. This seems “dark” (as Davidson puts it), and Frege thought so himself, though he thought it was the best we could do for natural languages. In an artificial language, different words should be used fore the two duties, thus relieving the overburdened verbs of the 1st, more obscure, duty. This leads, however, <i>ad infinitum</i> into an infinite primitive vocabulary.

Frege’s original proposal doesn’t work either, for even if we can make sense of (1), the task of reducing the determination of meanings to theory remains, which can be infinite.

A language proposed by A. Church meets Frege’s criteria for the language discussed above. New expressions gained by climbing the semantic ladder aren’t logically complex if constructed in the right way out of expressions of the next lower level. But this can’t be used as part of the theory of meaning. Church doesn’t dispute that each level’s expressions are semantical primitives, but Davidson submits that Church’s language is unlearnable.