Tarski and Truth

Setting the Stage The Liar Paradox Tarski’s Theory of Truth Setting the Stage The Liar Paradox
he Stage The Liar Paradox Tarski’s Theory of Truth
Announcements
For 10.29
Tarski’s Theory of Truth 1 Class on Monday 11.03 is cancelled

Liars, Languages & Logic 2 The syllabus has been updated accordingly (on Sakai)
3 There’s an extra-credit assignment buried in today’s
lecture
William Starr
It’ll increase your final grade by %1, if you get it right
10.29.08
William Starr — Tarski’s Theory of Truth (Phil 210.03) — Rutgers University 1/30 William Starr — Tarski’s Theory of Truth (Phil 210.03) — Rutgers University 2/30
Setting the Stage The Liar Paradox Tarski’s Theory of Truth Setting the Stage The Liar Paradox Tarski’s Theory of Truth
Outline Truth
It’s Everywhere
We’ve seen throughout this course that the notion of

truth is interconnected with that of meaning
1 Setting the Stage Frege took ‘The True’ & ‘The False’ as primitive,
undefinable notions
2 The Liar Paradox Sentences refer to The True or The False
He also said explicitly that Truth was undefinable
3 Tarski’s Theory of Truth (The Thought:37-8)
Tarski saw this as a challenge:
How could we define the word truth?
Can we define it without positing mysterious,
scientifically unrespectable entities like The True?
Tarski’s answer: Yes!
Truth What does True Mean?

What Does it Apply to? The Correspondence Theory
The Word True Applies to Sentences The Correspondence Theory of Truth

Tarski: let’s talk about truth applying to sentences The truth of a sentence consists in its correspondence to
reality
Sentences are concrete, scientifically respectable
entities Tarski wants to use this classical conception of truth as
Sets of sufficiently ‘similar’ inscriptions or his starting point
vocalizations It’s realistic, and seems to make clear sense
This makes the concept of truth relative to a language It’s derived from Aristotle
Since true applies to some specified set of linguistic
To say of what is that it is not, or of what is not
expressions, i.e. a language
that it is, is false, while to say of what is that it is,
Tarski : if the notion of proposition is ever clarified, or of what is not that it is not, is true.
maybe we can apply our results them instead (From Aristotle’s Metaphysics)
Defining Truth Truth

Give Me Something to Work With! Platitudes About Snow
Sentences like (1) are self-evident

We’ve said that true is supposed apply to sentences
(1) “Snow is white” is true if, and only if, snow is white
We’ve also said that truth is correspondence with
You would expect this, given the correspondence
reality
theory
Fine, but this doesn’t seem to provide us much help in
But, what exactly is a sentence ‘like’ (1)?
defining truth
Take a sentence S, and form a name for it by enclosing
To remedy this situation, Tarski isolated a key and
it in quotation marks: “S ”
uncontroversial aspect of truth
Now, what’s the logical relationship between S & (2)?
As it turns out, this aspect highlighted some very
tangible goals for a theory of truth (2) “S ” is true
Well, “S ” is true if, and only if, S
Schema (T) The Material Adequacy Condition

What it is, What it is for On Definitions of True
Schema (T) Tarski’s Material Adequacy Condition

We call the following Schema (T) For a definition of truth for a language L to be
(T) “S ” is true if, and only if, S adequate, it must entails every instance of (T)
We get an instance of Schema (T) by plugging in an (T) “S ” is trueL if, and only if, S
actual sentence for S (Where S is a sentence of L)
Example Instance: (T) itself is not a definition of truth
“Snow is white” is true if, and only if, Snow is white
Okay, so now we have a very concrete goal to meet in
Alright, what the %#*! is Schema (T) good for? defining true
Tarski : however you define truth, you should Isn’t it easy to meet?
guarantee that every instance of (T) is valid! Turns out: no
Problems for An Adequate Definition The Liar Paradox

The Liar Paradox What Just Happened
A Liar Sentence A Liar Sentence

(3) Sentence (3) is not true
(3) Sentence (3) is not true
(5) Sentence (3) is true iff sentence (3) is not true
An adequate definition of truth commits us to:
(T) “S ” is true if, and only if, S If we have an adequate definition of truth, according to
(4) “Sentence (3) is not true” is true iff sentence (3) is Tarski’s criterion, we are committed to (5)
not true That is, accepting every instance of Schema (T) entails
But: sentence (3) is “Sentence (3) is not true” the contradictory (5)
Since these sentences are identical, we can validly swap So, either Tarski needs to give up some instances of
them in (4) (by Leibniz’s Law) to get a contradiction: Schema (T), accept a contradiction or find some other
(5) Sentence (3) is true iff sentence (3) is not true
way of blocking this inference
Tarski goes for the last option
Languages & Logic Semantic Closure

Object Language vs. Metalanguage Defective Languages
When we consider instances of (T) we need to think

about languages A Liar Sentence
(T) “S ” is true if, and only if, S (3) Sentence (3) is not true
The S on the left belongs to the object-language
An instance of (T) itself belongs to the meta-language By allowing a definition of truth to apply to (3), we
have assumed (I)
(1) “Snow is white” is true if, and only if, snow is white (I) We can define truth for a language that itself
Object-language = Meta-language contains the predicate true, and the resources for
forming names of it’s sentences; i.e. quotation,
(6) “Schnee ist Weiss” is true if, and only if, snow is white labeling, etc. (Call such languages semantically closed)
Tarski suggests that we solve the paradox by giving up
Object language 6= Meta-language
this assumption
Note: translation of S must be used on right side of
biconditional
Giving Up (I) Tarski’s Limited Goal

What Does it Amount to? Where to go from here?
Using the lingo form a few slides back:

In defining truth for an object-language L1 , we must
employ a predicate trueL1 in the meta-language L2 So, Tarski doesn’t believe that we can fully capture our
If we give up (I), L1 can’t contain trueL1 ordinary notion of truth while remaining consistent
But L2 must have that predicate, so the Instead, he suggests that we theorize about truth in an
meta-language and object-language can’t be identical ever-ascending hierarchy of languages
Further, the meta-language must be strictly richer
In this way, we can maintain the core of our concept of
than the object-language
truth — Schema (T) — without lapsing into
An interesting consequence follows: inconsistency
Truth is undefinable for English (or any natural
Let us see how this kind of hierarchy allows one to
language), since English already contains its own
truth predicate define truth for a language
Tarski : that’s why we need an artificial language to
conduct science in!
A Truth Definition for wl A Truth Definition for wl

The Syntax The Semantics
I’ll describe a Tarskian truth definition for a simple Axioms (Reference & Application Axioms)
language
Wimpy Logic (wl)
1 P applies to an object o iff P is Ran and o ran
2 o is the referent of n iff either (i) or (ii) holds
First, we need to specify the syntax of wl
(i) n is jay and o is jay
The Syntax of wl (ii) n is kay and o is kay
1 Names: jay, kay
Definition (Inductive Definition of True wl )
2 Predicates: Ran(·)
1 P(n) is truewl iff P applies to the referent of n
3 If n is a name & P is a predicate, P(n) is a Wff
2 ¬A is truewl iff A is not truewl
4 If A is a Wff, ¬A is a Wff
3 A ∧ B is truewl iff both A and B are truewl
5 If A, B are Wff, (A ∧ B) is a Wff
6 Nothing else is a Wff The Meta-language here is ‘logical English’
Deriving Instances of (T) Extra Credit

Material Adequacy Do Some Derivations
Can we derive an instance of (T) for ¬Ran(jay)?

¬Ran(jay) is truewl iff Ran(jay) is not truewl (D2) Extra Credit
iff Ran does not apply to the referent of jay (D1) Complete a ‘step by step’ derivation of an instance of (T)
iff Ran does not apply to jay (A2.(i)) for the following sentences:
iff it’s not the case that Jay ran (A1) Ran(jay) ∧ Ran(kay)
¬(Ran(kay) ∧ Ran(jay))
We’ve got an instance of (T)!
¬Ran(jay) ∧ (Ran(kay) ∧ ¬Ran(jay))
Are you sure?
By ‘step-by-step’ I mean a line-by-line derivation, where
The correct translation of ¬Ran(jay) in logical English
each line is derived from the previous one by appealing to
is It’s not the case that Jay ran
exactly one definition or axiom, as stated on slide 25.
Important: Tarskian truth theories require taking the
notion of translation for granted
Deriving Instances of (T) Tarskian Truth Definitions

Material Adequacy Beyond Wimpy Logic
We’ve just demonstrated that we can derive an We saw a truth definition for a very simple language,
instance of (T) for ¬Ran(jay) but things get more complicated for richer languages
This is especially true for languages with quantifiers
But remember, for our truth definition to be materially
and variables
adequate we must show that we can derive every
If L contains quantifiers and variables, true L cannot
instance of (T)
be directly defined
There are infinitely many, so this requires some Instead trueL must be defined in terms of satisfaction
ingenuity
Tarski’s paper discusses it a bit, as does §§9.4, 18.4 of
But, it can be done, although we won’t attempt it in the LPL textbook on Sakai
this class
Consult them if you’re curious
Setting the Stage The Liar Paradox Tarski’s Theory of Truth
Reading for Next Class

Reading for 11.05
For Wednesday 11.05, read:

Mandatory Davidson’s “Truth & Meaning” (in
Martinich)
Ch.4 of Lepore & Ludwig’s Donald
Davidson: Meaning, Truth, Language, and
Reality (on Sakai)
Optional Ch.3 of Lepore & Ludwig’s Donald
Davidson: Meaning, Truth, Language, and
Reality (on Sakai)
Remember: class next Monday is cancelled
William Starr — Tarski’s Theory of Truth (Phil 210.03) — Rutgers University 30/30

Tarski and Truth

Uploaded by

Copyright:

Available Formats

Tarski and Truth

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tarski and Truth

Uploaded by

Copyright:

Available Formats

Setting the Stage The Liar Paradox Tarski’s Theory of Truth Setting the Stage The Liar Paradox

he Stage The Liar Paradox Tarski’s Theory of Truth

Tarski’s Theory of Truth 1 Class on Monday 11.03 is cancelled

We’ve seen throughout this course that the notion of

Truth What does True Mean?

The Word True Applies to Sentences The Correspondence Theory of Truth

Defining Truth Truth

Sentences like (1) are self-evident

Schema (T) The Material Adequacy Condition

Schema (T) Tarski’s Material Adequacy Condition

Problems for An Adequate Definition The Liar Paradox

A Liar Sentence A Liar Sentence

Languages & Logic Semantic Closure

When we consider instances of (T) we need to think

Giving Up (I) Tarski’s Limited Goal

Using the lingo form a few slides back:

A Truth Definition for wl A Truth Definition for wl

Deriving Instances of (T) Extra Credit

Can we derive an instance of (T) for ¬Ran(jay)?

Deriving Instances of (T) Tarskian Truth Definitions

Setting the Stage The Liar Paradox Tarski’s Theory of Truth

Reading for Next Class

For Wednesday 11.05, read:

You might also like