The main point of Peter Geach's paper, "Assertion" (Logic Matters, Basil Blackwell, 1972, pp. 254-269) is what he calls the Frege point: A thought may have just the same content whether you assent to its truth or not; a proposition may occur in discourse now asserted, now unasserted; and yet be recognizably the same proposition. This seems unassailably correct. One will fail to get the Frege point, however, if one confuses statements and propositions. An unstated statement is a contradiction in terms, but an unasserted proposition is not. The need for unasserted propositions can be seen from the fact that many of our compound assertions (a compound assertion being one whose content is propositionally compound) have components that are unasserted.
To assert a conditional, for example, is not to assert its antecedent or its consequent. If I assert that if Tom is drunk, then he is unfit to drive, I do not thereby assert that he is drunk, nor do I assert that he is unfit to drive. I assert a compound proposition the components of which I do not assert. The same goes for disjunctive propositions. To assert a disjunction is not to assert its disjuncts. Neither propositional component of Either Tom is sober or he is unfit to drive is asserted by one who merely asserts the compound disjunctive proposition.
What bearing does this have on recent discussions? I am not sure I understand William of Woking's position, but he seems to be denying something that Geach plausibly maintains, namely, that "there is no expression in ordinary language that regularly conveys assertoric force." (261) Suppose I want to assert that Tom is drunk. Then I would use the indicative sentence 'Tom is drunk.' But there is nothing intrinsically assertoric about that sentence. If there were, then prefixing 'if' to it would not remove its assertoric force as it does. As I have already explained, an assertive utterance of 'If Tom is drunk, then he is unfit to drive' does not amount to an assertive utterance of 'Tom is drunk.' 'If' cancels the assertoric force. And yet the same proposition occurs in both assertions, the assertion that Tom is drunk and the assertion that if Tom is drunk, then he is unfit to drive. I conclude that there is nothing intrinsically assertoric about indicative sentences. If so, there is no semantic component of an indicative sentence that can be called the assertoric component.
'If' prefixed to an indicative sentence does not alter its content: it neither augments it nor diminishes it. But it does subtract assertoric force. Given that the meaning of an indicative sentence is its content, and the semantics has to do with meaning, then there is no semantic assertoric component of an indicative sentence or of the proposition it expresses. Assertion and assertoric force do not belong in semantics; they belong in pragmatics. Or so it seems to me.
You have made these points before, but I will reply again.
>>A thought may have just the same content whether you assent to its truth or not; a proposition may occur in discourse now asserted, now unasserted; and yet be recognizably the same proposition. This seems unassailably correct.
It depends how you define 'thought'. I have claimed below that the semantics of a sentence includes a 'content', plus an assertoric component. The semantic value of the 'content' corresponds to the noun-clause formed by prefixing the 'that' operator to the sentence in question. If the sentence (i.e. the sentence-type) is 'Tom runs', the noun-clause is
that Tom runs
This is probably what Geach means by 'thought'. I.e. I can think that Tom runs, wonder whether Tom runs and so on. So the 'thought', if it is the same as my 'content', is not equivalent to the full semantics of the sentence (as I claim) because it does not include the assertoric component.
On the claim that "a proposition may occur in discourse now asserted, now unasserted; and yet be recognizably the same proposition" I don't have the definition of 'proposition' here. Does it mean sentence? Does it mean thought/content?
>>One will fail to get the Frege point, however, if one confuses statements and propositions. An unstated statement is a contradiction in terms, but an unasserted proposition is not.
I hold there are sentences (sentence-types), and meanings.
>>The need for unasserted propositions can be seen from the fact that many of our compound assertions (a compound assertion being one whose content is propositionally compound) have components that are unasserted.
The examples you then give, you have given before, and are adequately explained by the concept of 'semantic cancellation'. See my post here http://ocham.blogspot.com/2010/05/semantic-cancellation.html explaining this.
>>I am not sure I understand William of Woking's position, but he seems to be denying something that Geach plausibly maintains, namely, that "there is no expression in ordinary language that regularly conveys assertoric force."
It should be plain that I am not denying this. If I held a compositional theory according to which the semantic composition of the sentence exactly corresponded to the lexical composition, then yes. But I have explained a few times my claim that these are different. The continuing difficulties you are having with this are caused by your failure to appreciate this point.
To further clarify. Let E be any expression type (e.g. 'Tom runs').
1. Assume there is a function m that maps expression types onto a meaning. Thus, m(E) is a meaning.
2. Assume that a meaning can be composite, i.e. consists of parts or components. Define 'a is in m(E)' as meaning that meaning a is part of the meaning m(E).
3. Semantic cancellation: if m(C+E) is in m(E), and not m(C+E) = m(E), then we say that the addition of expression C cancels some of the meaning of E. Given compositionality (2 above) it follows that there is some component of the meaning of E - the cancelled component - that is present in E but not in C+E.
>>Suppose I want to assert that Tom is drunk. Then I would use the indicative sentence 'Tom is drunk.' But there is nothing intrinsically assertoric about that sentence.
Yes there is. Compare this to a proton. A proton is a type of particle that intrinsically has a positive charge. Any token of that type has a positive charge. Equally, any token of 'Tom runs' contains an assertoric component.
>>I have already explained, an assertive utterance of 'If Tom is drunk, then he is unfit to drive' does not amount to an assertive utterance of 'Tom is drunk.' 'If' cancels the assertoric force.
Yes, and you have just said that 'if', a semantic cancellation operator, cancels the force. Compare, when you add an electron to a proton, the net charge is zero. This does not mean that the proton loses its charge. Not at all. It has to maintain its charge, in order for the net charge to be zero.
>>And yet the same proposition occurs in both assertions, the assertion that Tom is drunk and the assertion that if Tom is drunk, then he is unfit to drive. I conclude that there is nothing intrinsically assertoric about indicative sentences. If so, there is no semantic component of an indicative sentence that can be called the assertoric component.
The same sentence occurs in both sentences. You might as well argue that, because the same type of particle (a proton) occurs in an ion (which has a charge) and an atom (which has no charge), that there is nothing intrinsically positive about a proton.
>>'If' prefixed to an indicative sentence does not alter its content: it neither augments it nor diminishes it.
Correct, just as adding an electron to a proton neither augments nor diminishes the charge of the proton.
>>But it does subtract assertoric force.
In the sense that adding an electron reduces the net charge of the composite proton+electron.
>>Given that the meaning of an indicative sentence is its content, and the semantics has to do with meaning ...
No, the content of an indicative sentence is somewhat less than its semantics. The semantics includes the content+assertoric component.
>>... then there is no semantic assertoric component of an indicative sentence or of the proposition it expresses.
If meaning (semantics) = content, then yes. But it is patently clear that what the noun-clause 'that Tom runs' signifies is somewhat less than the semantics of 'Tom runs'.
>>Assertion and assertoric force do not belong in semantics; they belong in pragmatics. Or so it seems to me.
I'm not sure what 'pragmatics' really means. Let's distinguish between lexical meaning, which belongs to expression-types, and anything else. Then I claim that the expression-type 'Tom runs' has the same semantics wherever it occurs (content + assertoric force), just as a proton has the same charge, however tokened by any particular particle. When such a particle is combined with others, however, the aggregate will have a different charge. So too with semantics.
Posted by: William | Thursday, June 03, 2010 at 04:30 AM
Bill, looking at my reply above I need to ask a favour. A favour, because in replying you are helping me put together arguments, and clarify matters that need clarifying and generally help me to get things in a state of publication, if necessary, for which I am indebted.
It would help me a lot if (a) where an argument of mine is clear, to reply specifically to that argument. Some of the time (as above) you are simply repeating arguments you have made before, and to which I reply exactly the same as before. This does not get us far. If you could 'reply to the replies' it would make it easier. (b) where you don't find an argument or a point clear, to say exactly where it is unclear. Saying or implying or suggesting that my whole position is unclear does not help me address the locus of unclarity.
The main points I have been making, which I hope are clear, are
1. That the semantics of a sentence can have components. E.g. if the meaning is M, then we can intelligibly say that some a is a component part of M.
2. That the semantic components of a sentence are (or can be) distinct from the lexical components. This is a really important point that addresses (in my view) most of your objections.
3. That the semantics of a sentence is a function of the lexical composition. The idea of a function is one that is fundamental to mathematics, and which I hope is clear for the present purpose.
4. That the semantics is a function of sentence-type, i.e. a token of any sentence-type has the same semantics, however betokened.
If any of these points aren't clear, please say. If they are incorrect, then let's look at why they are incorrect, i.e. reasons.
William
Posted by: William | Thursday, June 03, 2010 at 05:59 AM
I'll try to be helpful. If we could sit down with each other, face to face, and had many hours we could probably iron out our terminology -- which I think is a major part of our difficulty.
Ad (1). I have no problem with compositionality of meaning, but what exactly do you mean by 'meaning' as applied to sentences? Do 'One dies' and 'Man stirbt' have the same meaning? Until we know what all is included and not included in your concept of meaning we will not know what all is included in your concept of semantics. For example, 'Der Mensch ist ewig' has a meaning, but it is false. Are truth and falsity semantic concepts?
Ad (2). Semantic and lexical. Take 'fishmonger' and 'spendthrift.' Both are lexically simple, but only the second is semantically simple. The first is semantically complex in that the meaning of 'fishmonger' is built up out of the meanings of 'fish' and 'monger.' Would you accept that?
Ad (3). You are saying that the meaning of a sentence is uniquely determined by its lexical composition. Can you give examples?
Ad (4). You are saying that if a sentence type has meaning M, then every token of this type has M. Consider 'We don't do that sort of thing.' One can easily imagine circumstances in which tokening that sentence type would amount to giving a command or an admonition and other circumstances in which a tokening of the same sentence type would be a mere declaration of fact. This motivates a distinction between sentence meaning and speaker's meaning. And this bring us back to exactly what 'meaning' means in these discussions, and what all is included in the scope of semantics.
Posted by: Bill Vallicella | Thursday, June 03, 2010 at 01:45 PM
>>Suppose I want to assert that Tom is drunk. Then I would use the indicative sentence 'Tom is drunk.' But there is nothing intrinsically assertoric about that sentence.
Yes there is. Compare this to a proton. A proton is a type of particle that intrinsically has a positive charge. Any token of that type has a positive charge. Equally, any token of 'Tom runs' contains an assertoric component.
I reject the analogy. 'If' cancels the assertoric force of 'Tom runs.' So there cannot be anything intrinsically assertoric about the sentence. In any case, I don't believe you have ever explained what 'assertoric' means. Do you just mean that the sentence is in the indicative mood?
Posted by: Bill Vallicella | Thursday, June 03, 2010 at 07:14 PM
Hi Bill, and thanks. Replying to your questions, but note we need to separate the question of whether a hypothesis is coherent or intelligible, from the question of whether it is true.
(1). Your question was about what I meant by 'meaning', and in particular, whether 'one dies' and 'Man stirbt' have the same meaning. Yes they do, on the assumption that 'Man stirbt' states that man dies (i.e. they have the same content, and they both assert that content).
(2). You say 'fishmonger' is semantically complex in that the meaning of 'fishmonger' is built up out of the meanings of 'fish' and 'monger.' Yes, I think I accept that. Aristotle gives a similar pair of examples (where one word is semantically complex, another is not).
(3). "You are saying that the meaning of a sentence is uniquely determined by its lexical composition. Can you give examples?" This was simply a counter to your claim that some tokens of the same type may have a different meaning to others.
(4) "One can easily imagine circumstances in which tokening that sentence type would amount to giving a command or an admonition and other circumstances in which a tokening of the same sentence type would be a mere declaration of fact. " - As I have argued, the phenomenon of semantic cancellation explains this. To be very precise, I hold that the same token will always have the same meaning, just as any instance of a proton always has a positive charge. However, the token may occur as part of a larger aggregate where its effect is cancelled, just as a proton may be found with an electron, which has a negative charge.
(5) In your second comment you reject the analogy between semantic forces and electrical charge. OK but the reason? Your argument was that if the word 'if' cancels the assertoric force of 'Tom runs.', there cannot be anything intrinsically assertoric about the sentence. Why? Why is this reason valid in this case, but not valid in the case of electrons? 'The electron cancels the charge of a proton, so there cannot be any intrinsic charge of a proton' is not valid. Nor is 'the weight of Tom on the see-saw cancels out the weight of Daisy on the other side, so Tom cannot have any intrinsic weight'. Invalid.
(6) You say I haven't explained what 'assertoric' means. Do I just mean that the sentence is in the indicative mood? To the first, I did explain this in a comment to an earlier post, some time ago. To the second, no, I don't mean just 'indicative mood'. I define 'assertoric component' as follows. I hold (A) that there is a difference in meaning between the expression type 'Tom runs', which is a declarative sentence, as it happens, and the expression type 'that Tom runs', which is a noun phrase. (B) difference in meaning I define as the assertoric component. Essentially, given that the (B) is a definition - de definitionibus non est disputandum, the dispute is whether 'Tom runs' and 'that Tom runs' have the same meaning or not. Do you think they have the same meaning?
Posted by: William | Friday, June 04, 2010 at 03:11 AM
Sorry I should of course have written 'Man stirbt' states that one dies.
Posted by: William | Friday, June 04, 2010 at 06:08 AM
If I've got him right WW is saying that the meaning of sentence s="if Tom is drunk then Tom is unfit" is a function m of the separate meanings of sentences s1="Tom is drunk" and sentence s2= "Tom is unfit". Also that a meaning is a 2d vector, which I'll write as [c,a], the first component being a 'content' and the second an 'assertoric level'. Assertoric level is just a boolean 1 or 0 indicating the presence or absence of assertion, respectively.
m(s) = if_then_ (m(s1), m(s2))
= if_then ( [c (s1), a(s1)] , [c(s2), a(s2)])
= [implies(c(s1), c(s2), 1]
In other words, if_then_ simply discards the assertoric components of its two argument vectors and returns a meaning vector with second component 1. This all makes good sense in terms of the composition of functions. Are metaphors drawn from physics helping here?
Posted by: David Brightly | Friday, June 04, 2010 at 06:48 AM
DB: >>a function m of the separate meanings of sentences
The function I had in mind would always be from sentence-tokens to meanings. If it is from meaning to meaning, as here
if_then ( [c (s1), a(s1)] , [c(s2), a(s2)])
you have the problem that the meaning [c (s1), a(s1)], for example is 'still there'. And if it is 'still there', how isn't 'Tom is drunk' still being asserted? That is fundamentally Bill's objection.
>>Are metaphors drawn from physics helping here?
I think we need the metaphor to explain the cancellation effect. On the account I am giving, the assertion of 'Tom is drunk' is still present in
(*) If Tom is drunk, then Tom is unfit
but it is being cancelled out by the 'if then' part, in the way that the electron cancels the charge of the proton. The charge of the electron cancels out the charge of the proton not by removing it, but by counterbalancing it. The weight of Tom in the seesaw does not cancel out the weight of Daisy, but balances it.
Posted by: William | Friday, June 04, 2010 at 09:05 AM
To be clear about Bill's argument (as I understand it). Bill is pointing out the logical inconsistency of the following three claims.
(1) Every token of the same expression type has the same meaning
(2) Assertion is a semantic component of a declarative sentence (i.e. sentence type)
(3) Some declarative sentence tokens (e.g. the 'Tom is drunk' in 'if Tom is drunk, Tom is unfit') do not carry an assertoric component.
I hold both (1) and (2). I hold (1) because I claim that assertion is part of 'linguistic meaning', and 'linguistic meaning' essentially belongs to expression types. I hold (2) because that is fundamental to the theory of truth I am proposing.
However, I also deny (3). The 'Tom is drunk' in 'if Tom is drunk, Tom is unfit' does carry assertoric force. The force is still 'there'. However, because it is part of a larger sentence which has a cancellation component, the force is cancelled out. Cancellation here means not that it is annihilated, but that its aggregated effect is nullified by another equal but opposite effect. Much the same way that when two children are sitting on a seesaw, they retain their original weight, and each presses down with the same force as if only one were sitting. But when both are equal weight, and both are sitting on the seesaw, no net movement results.
Posted by: William | Friday, June 04, 2010 at 09:18 AM
W writes, >>(1). Your question was about what I meant by 'meaning', and in particular, whether 'one dies' and 'Man stirbt' have the same meaning. Yes they do, on the assumption that 'Man stirbt' states that man dies (i.e. they have the same content, and they both assert that content).<<
It is clear that they have the same meaning since each is a translation of the other. This meaning is as it were assigned to them by the language systems of which each is a part, English and German. So I say they have the same linguistic meaning. But I object when you say that either sentence states or asserts anything. Neither, by itself, does any such thing. Only persons state or assert. Is that not a plain fact? If yes, then it is not clear what an assertoric semantic component is or could be.
Posted by: Bill Vallicella | Friday, June 04, 2010 at 12:25 PM
You maintain that there is never a situation in which two tokens of the same sentence type differ in meaning. I disagree. Consider 'Notice that Joan is wearing her purple hat again.' In one situation a speaker produces a token of this type to give a command; in another to state a fact. The tokens differ in meaning despite being tokens of the same type.
Posted by: Bill Vallicella | Friday, June 04, 2010 at 02:27 PM
>>I hold (A) that there is a difference in meaning between the expression type 'Tom runs', which is a declarative sentence, as it happens, and the expression type 'that Tom runs', which is a noun phrase. (B) difference in meaning I define as the assertoric component. Essentially, given that the (B) is a definition - de definitionibus non est disputandum, the dispute is whether 'Tom runs' and 'that Tom runs' have the same meaning or not. Do you think they have the same meaning?<<
Your argument seems to be this:
1. There is a difference in meaning between 'Tom runs' and 'that Tom runs.'
2. There must be a meaning-element (semantic element) that accounts for this difference.
3. Define this semantic element as the assertoric component.
4. The assertoric component is a proper part of the meaning of 'Tom runs.'
Therefore
5. The meaning of 'Tom runs' has as a proper part a semantic element called the assertoric component which distinguishes this meaning from the meaning of 'that Tom runs.'
I grant (1). But (2) is not self-evident. Why can't the difference be a brute difference? Why the need for something to account for it? As for (3) you can use words any what you like, but since sentence types don't assert anything, your semantic component is mislabelled 'assertoric.' As for (4), why couldn't the assertoric component be an improper part of the meaning of 'Tom runs'?
Posted by: Bill Vallicella | Friday, June 04, 2010 at 02:57 PM
>> since sentence types don't assert anything
Evidence for this? The arguments you have given so far are vulnerable to the 'seesaw' fallacy (see above). Perhaps you mean that the dictionary definition of 'assert' requires that a human be the subject of this verb? But common usage (try Google) suggests otherwise.
>>There must be a meaning-element (semantic element) that accounts for this difference.
>>But (2) is not self-evident. Why can't the difference be a brute difference?
I don't understand what a 'brute difference' is. If you agree that
(1) Tom runs
differs in meaning from
(2) that Tom runs
and if you agree that
(3) It is true that Tom runs
has the same meaning as (1), then you agree to an analysis of (1) which is lexically complex in the way that I argue is semantically complex. I.e. 'it is true' corresponds to the 'assertoric' component, and 'that Tom runs' to the 'content' component.
>>As for (3) you can use words any what you like, but since sentence types don't assert anything, your semantic component is mislabelled 'assertoric.'
I think 'assertoric' is a good word, because 'assertoric' is traditionally used to signify the element which brings in truth or falsity. Only an assertion can be characterised by terms 'true' or 'false'. Since (I claim) the assertoric component corresponds to the expression 'it is true' (but not 'is is true that' - this is an important difference between my account and the redundancy theory) I think 'assertoric' is a fair choice.
Posted by: William | Saturday, June 05, 2010 at 12:05 AM
Second thoughts. On the principle that (as well as never conceding more than is necessary) one should concede as much as is consistent with one's ultimate position: I am happy to allow that the verb 'assert', and the verb 'state' cannot take an inanimate subject, but only a rational, conscious one, a human. Even though I do not agree with it, for other reasons (ordinary usage).
However, I continue to hold that it is only through signs (in particular, sign types) that we are able to signal statements or assertions. Thus in
(A) Is Tom running ? - yes.
the utterer is signalling, via the linguistic meaning of 'yes', that he or she is asserting something that admits of truth or falsity. Similarly the statement
(B) The proposition that Tom runs .... is true
contains two distinct (lexical) elements, the bit which specifies the content (the that-clause) and the bit at the end 'is true' which signals assertion. And now I define 'assertoric component': that part of the sentence which signals that the speaker is making an assertion (and not just wondering, or entertaining, or questioning).
If you concede that such a signal is given in (B), and if you accept my definition of 'assertoric component' (namely the signal) then equally you must concede that the same signal is given, albeit implicitly, in
(C) Tom runs.
Posted by: William | Saturday, June 05, 2010 at 03:50 AM
WW,
As I understand it you see the assertoric charge of a sentence as in some sense the sum of the assertoric charges of its parts, in much the same way as the net charge of an assemblage of ions is the sum of its component ionic charges. This is still a functional relationship between the component charges and the resultant charge. But I'm not sure this simple additivity is sustainable.
AC(Tom runs)=1 (assertion)
AC(Tom runs?)=0 (question) Hence AC(?)=-1
AC(Tom runs? Yes)=1 (assertion again) Hence AC(Yes)=1
Now consider 'Tom runs. Yes?' Adding up the component assertoric charges we get a total of 1. But arguably this sentence is a question not an assertion and so should have assertoric charge 0. This suggests that assertoric charges do not combine commutatively.
Posted by: David Brightly | Saturday, June 05, 2010 at 07:38 AM
Hi David. Yes, the question mark clearly has 'charge' of minus 1. Adn the charge of 'yes' is plus 1, clearly. Your problem example is
Tom runs. Yes?
Surely we either read this as
1. A complete declarative sentence, stating that Tom runs, followed by a mere interjection 'yes', meaning 'do you agree', which is a question.
2. A combination of 'tom runs' and 'yes' to form an imperative - 'Tom is going running right now'. In this case, 'Tom runs' has a different semantics. Bill will probably say here that this supports his case, to which I reply that this reading requires ignoring the full stop (US - 'period') after 'runs'. If the words are spoken where you can't always hear the full stop), it amounts to simple equivocation.
3. A combination of 'tom runs' and 'yes' to form a question. Again, this requires a misreading of the full stop.
Posted by: William | Saturday, June 05, 2010 at 08:49 AM
[>> since sentence types don't assert anything
Evidence for this?]
"Tom is running" or "Tom runs" - used as a code for an operative, means tonight you are to put the poison in the ambassador's soup. This is a command, not an assertion.
The sentence need not actually assert some person Tom is running or assert anything for that matter.
"Equally, any token of 'Tom runs' contains an assertoric component."
This is simply not true. It is not the case that all syntactic sentences are sufficient for assertion, (nor is it necessary.)
Posted by: T. Hanson | Saturday, June 05, 2010 at 08:55 AM
W,
Ordinary usage counts for little in my book. People say the damndest things. Appeals to OL won't budge me.
>>(B) The proposition that Tom runs .... is true
contains two distinct (lexical) elements, the bit which specifies the content (the that-clause) and the bit at the end 'is true' which signals assertion. And now I define 'assertoric component': that part of the sentence which signals that the speaker is making an assertion (and not just wondering, or entertaining, or questioning).
If you concede that such a signal is given in (B), and if you accept my definition of 'assertoric component' (namely the signal) then equally you must concede that the same signal is given, albeit implicitly, in
(C) Tom runs.<<
My difficulty with this is your claim that 'is true' signals assertion. I say this because 'is true' can occur in an unasserted sentence. Consider this argument:
If it is true that Tom is drunk, then Tom is unfit to drive
It is true that Tom is drunk
Ergo
Tom is unfit to drive.
If a person were to give this argument, he would be asserting the minor but not asserting the protasis of the major. Now the minor and the protasis of the major have exactly the same sense. Therefore, assertion is no part of this sense.
In other words, 'It is true that p' can occur asserted or unasserted without any change of meaning. So assertion cannot be a part of the meaning: there is no assertoric component that is part of the semantic content.
Whatever assertoric force is, it cannot be identified with the meaning of 'it is true' or 'is true.'
Posted by: Bill Vallicella | Saturday, June 05, 2010 at 03:39 PM
T. Hanson,
I think William has the resources to turn aside your objection. One could set up a code in which 'Tom runs' means poison the ambassador. To evade this, W. could refine his thesis by saying that he is concerned with sentence types as carrying their conventional meanings as assigned by the language systems to which they belong.
In other words, he is not talking about physical marks (with their geometrical and color properties)but about such marks as embodying a conventional meaning.
Posted by: Bill Vallicella | Saturday, June 05, 2010 at 03:48 PM
WW,
Well, if we allow "Tom runs? Yes" to count as a single assertion I think we should allow "Tom runs. Yes?" to count as a single question. But perhaps we shouldn't press you too hard with these edge cases from ordinary usage. Let's accept that the assertoric charge metaphor gives the right answer for the assertoric status of an if_then_ sentence. What it doesn't really explain is how the charge is distributed. Haven't you got to say something like this: the assertoric subsentences "Tom is drunk" and "Tom is unfit" are both like a -1 negative ion, ie they are neutral atoms (contents) that have gained an electron (assertion). The 'if_then_' operator is like a +1 hydrogen ion, ie a bare proton. In combination, in the sentence "if Tom is drunk then Tom is unfit", the if_then_ hydrogen ion captures the two electrons from the subsentence ions, rendering them neutral, ie non-assertoric content, and the compound becoming negative, ie assertoric. In other words, the contents of the two subsentences are bound together by something representing implication and the whole given assertoric status. This is exactly what is conveyed by the functional specification I gave earlier. One might fancifully say that just as chemistry is about the movement of electronic charge around molecules (compounded atoms), so assertion semantics is about the movement of assertoric charge around compounded subsentences.
Posted by: David Brightly | Saturday, June 05, 2010 at 07:31 PM
>>My difficulty with this is your claim that 'is true' signals assertion. I say this because 'is true' can occur in an unasserted sentence. Consider this argument:
If it is true that Tom is drunk, then Tom is unfit to drive ...
There was a rule of scholastic debate that the first person to repeat themselves loses the argument. Actually I have been repeating myself for some time, so have you.
I have replied several times to your 'if then..' argument that this is explained by 'semantic cancellation'. I even resurrected my blog to talk about this. The word 'that' cancels the assertoric force of 'Tom runs' ('cancels' in the sense that an electron cancels the charge of a proton, Daisy cancels Tom weight on the seesaw, even though the proton still has charge, and Tom still has weight)
The words 'it is true' put the force back. And we can even cancel this again, as in
(*) that it is true that Tom runs
and so on. So also for 'if ... then' statements. Perhaps we should put this on hold for a time. (Bill, note my email about the IEP article, which would be a breath of fresh air.
>> 'It is true that p' can occur asserted or unasserted without any change of meaning.
Likewise, Tom can be sitting on the seesaw and force it to the ground (without Daisy on the other side), or he can fail to force it to the ground (with Daisy opposite). In both cases, Tom has weight. Likewise 'It is true that p' can occur on its own, or in a more complex sentence. In both cases it retains assertoric force, just as in both cases Tom has weight. However, when included in a more complex sentence, it may lose its force in the whole (just as Tom's weight fails to push down the seesaw when Daisy is on the other side).
Posted by: William | Sunday, June 06, 2010 at 12:33 AM
Apologies my comment above seems to have come through twice.
Posted by: William | Sunday, June 06, 2010 at 12:35 AM
"To evade this, W. could refine his thesis by saying that he is concerned with sentence types as carrying their conventional meanings as assigned by the language systems to which they belong."
In that case, I forgot the most important part of the story ;-)
Meanwhile the ambassador is dressing for the fateful state dinner, and procrastinating. His wife says impatiently "time is ticking."
Is this an assertion out of the blue the that time has a direction?
More likely it is simply a command: "hurry up!"
Posted by: T. Hanson | Sunday, June 06, 2010 at 08:39 AM
W,
I say 'is true' has nothing to do with assertion. You deny this. I am not sure how we can move beyond this standoff.
I see if that IEP article sheds any light. Do you endorse its content?
Posted by: Bill Vallicella | Sunday, June 06, 2010 at 12:46 PM
>>I say 'is true' has nothing to do with assertion. You deny this. I am not sure how we can move beyond this standoff.
The standard way is to provide arguments and reasons, rather than blunt assertions. You gave an argument several times above (the 'if then' argument). I replied to the argument several times. You have not replied to this reply. To move beyond the standoff, you would need to reply to the reply.
On the IEP article, I have commented here http://ocham.blogspot.com/2010/06/prosentential-theory-of-truth.html
Posted by: William | Sunday, June 06, 2010 at 11:28 PM
@Hanson
Your point can easily be explained by "enthymeme" - an argument where one or more of the assumptions is so obvious that it need not be stated. Suppose at a party my wife says "It's past eleven". I know what she means, which can be expressed by the following syllogism.
(Major) If it is past eleven and we are out, then leave now!
(Minor) It is past eleven and we are out
(Conclusion) Leave now!
She does not need to state the major premiss, nor even the conclusion, but I catch her drift. Note both the major and the conclusion contain an imperative. Arthur Prior investigated the logic of arguments containing imperatives in his paper "On Some Proofs of the Existence of God" , published in Papers in Logic and Ethics Duckworth 1976.
Note also that the minor premiss is an assertion, and must be understood as such for the argument to be valid. If by contrast it meant something like "Would you like another large gin and tonic?" her conclusion would not follow. The point being, "linguistic meaning" is king. If it were not for the constant meaning we assign by convention to our common language, we would not be able to express ourselves in ways like this, that superficially suggest that meaning constantly fluctuates, and that some expressions have no linguistic meaning.
Posted by: William | Monday, June 07, 2010 at 12:54 AM
Bill's rhinotillexomania example can also be seen as enthymematic. Mommy says
expecting Johnny to deduce and hoping that Johnny's desire to be one of us exceeds his fondness for doing that.Posted by: David Brightly | Monday, June 07, 2010 at 01:44 AM
W,
The argument I gave struck me as well-nigh decisive. In any case, your replies are merely restatements of what you've been saying all along. Once I finish studying the IEP Prosentential Theory of Truth entry along with the post on your blog, I'll start another round. Perhaps we can then ascend to higher ground whence to reconnoiter our skirmishing.
I suspect the outcome is fated: I shall remain a wild-eyed anti-deflationist with respect to existence and truth, and you the opposite. Our arguments will cancel each other.
Posted by: Bill Vallicella | Monday, June 07, 2010 at 12:13 PM
>>The argument I gave struck me as well-nigh decisive. In any case, your replies are merely restatements of what you've been saying all along.
But you still need to 'reply to the reply', which you haven't. But let's wait for the IEP article and have another round.
I thought your post on Nietszche above was intriguing.
Posted by: William | Tuesday, June 08, 2010 at 03:00 AM
>>In any case, your replies are merely restatements of what you've been saying all along.
Ingenious! you claim that my thesis was followed by your objection, which was followed by a restatement of my thesis. If that were true, by the rules of this chess-game I lose. But that is not a correct representation of the argument so far. See my summary below. You need to reply to my reply, namely my 'disproof of the minor (minor disprobatur). You need either to show that semantic cancellation does not exist, or is incoherent, or that it exists and is coherent, but does not apply in this case. None of what you have argued so far even mentions the concept 'semantic cancellation'.
Logical argument is just like a chess game. We have a common understanding of the rules of inference. The game ends either in reaching disagreement about a principle that is demonstrably fundamental, i.e. it self-evidently admits of no proof or disproof (Bill hates carrots), in which case stalemate, or where both sides end in agreeing a set of fundamental principles from which the truth of the winner's thesis follows with logical certainty.
---------------------- The argument so far -------------------------
(Woking Thesis) Expression types (e.g. a declarative sentences) can have assertoric force
[Vallicella objection]
(Major) If an expression-type has assertoric force, every token of it has assertoric force
(Minor) A token of any sentence may occur in a context where it has no assertoric force
(Conclusion) No expression-type has assertoric force.
(Proof of the minor) Take any declarative sentence-type such as 'Socrates runs'. But it has no assertoric force in the consequence 'If Socrates runs, Socrates moves'.
(Reply to objection)
I concede the argument of the objection is valid.I concede the major. I dispute the minor. Against the proof of the minor. 'Socrates runs' does have assertoric force in the 'If Socrates runs, Socrates moves'. However, its force is cancelled out by the 'if then' operator.
Posted by: William | Tuesday, June 08, 2010 at 03:51 AM
William writes:
"The point being, "linguistic meaning" is king. If it were not for the constant meaning we assign by convention to our common language, we would not be able to express ourselves in ways like this, that superficially suggest that meaning constantly fluctuates, and that some expressions have no linguistic meaning."
The consequent of your major premise can change depending on context. It could be "leave now." Or it could be "let's start our jumping jacks." There is no formal element of "It is eleven" itself that makes it an assertion, or a particular assertion about the time. A physical gesture could be an assertion. So I see no reason to say linguistic meaning is king. Meanings depend on implicit understandings of the kind expressed in your major premise - how symbols are used within particular social contexts. Understanding these particular meanings expressed in your major depends on social intelligence and sometimes subtle understandings of "forms of life" to use Wittgenstein's phrase. Language does fluctuate, that is why bad can mean good in some "forms of life." The fact that meaning shifts in background social contexts does not entail the expressions are meaningless, though it could. A gangster would find much of this dicussion meaningless, and we woud find many of his expressions meaningless.
Posted by: T. Hanson | Thursday, June 10, 2010 at 09:09 AM