Is free will compatible with determinism?

David W. Agler

The Pennsylvania State University

Incompatibilism

Goals and agenda

Last time: impossibilism (free will is impossible).
This time: incompatibilism (free will and determinism are inconsistent)
Goals
1. Define incompatibilism
2. Examine the primary argument for incompatibilism
3. Decide if we are incompatibilists

Incompatibilism defined

Incompatibilism

Determinism and the free will thesis are inconsistent.

If the world is deterministic, then we are not free.
If incompatibilism is false, then there is at least one possible world where (1) every event (besides the first one) is determined by the prior state of affairs and the laws of nature (the world is determined) and (2) at least one event is free.

Incompatibilism visualized

flowchart LR
    L[Laws]
    W1[World at time t1]
    W2[World at time t2]
    L & W1 --> W2

flowchart LR
    L[Neurological laws]
    W1[Brain state at t1]
    W2[Brain state at t2]
    A[Action at t3]
    L & W1 --> W2
    W2 --> A

Incompatibilism vs impossibilism

If the world is deterministic, then there is no free will.
If the world is indeterministic, then there is no free will.

Impossibilism accepts (1) and (2). Incompatibilism is only committed to (1).

Incompatibilists

Many incompatibilists think free will and indeterminism are consistent.
But, it is not necessary for you to accept this.

Most philosophers (n=1758) reject or lean toward a rejection of incompatibilism.

The Argument for Incompatibilism

The consequence argument

The primary argument for incompatibilism is known as the consequence argument.
Conclusion: if determinism is the case, then we do not have the capacity to do otherwise.
We don’t need to prove determinism

I’m going to give you a simplified version.

Step 1: Assumption

Assume determinism is true.
Necessarily, ((Laws + World at \(t_i\)) \(\to\) World at \(t_{i+j}\))
Now let’s show that if determinism is true, then there is no free will.

Step 2: Controlling entailments

No one has (or ever had) any choice or control about the fact that the (Laws + World at \(t_i\)) \(\to\) World at \(t_{i+j}\)

You can’t control the fact that the laws + past world entail later states in the world.

Step 3: Controlling laws

We can’t control the laws of nature.

Newton’s second law is that the acceleration of an object depends directly on the net force acting on it and inversely on its mass (\(F=ma\)).
Neither you or I or anyone has (or ever had) any choice or control over this law.

Step 4: Controlling laws

No one has (or ever had) any choice or control about the past state of the world.

We can’t control the past.
Example: Can’t change that dinosaurs existed.
Example: Can’t change that I ate a sandwich yesterday.

Step 5: Consequence argument

P1: Necessarily, (Laws + World at \(t_i\)) \(\to\) World at \(t_{i+j}\)
P2: No one has control (Laws + World at \(t_i\)) \(\to\) World at \(t_{i+j}\)
P3: No control about the Laws
P4: No control about the Past.
IC: Therefore, no control about the World at \(t_{i+j}\)
C: Therefore, no free will.

Exercise - Keeping it simple

Exercise

Try to express the Consequence Argument in the crudest, simplest, plainest way you can:

You can use slang
You can use a diagram or picture
Use a real life example if it helps

Criticisms

P1: Necessarily, (Laws + World at \(t_i\)) \(\to\) World at \(t_{i+j}\)
P2: No one has control (Laws + World at \(t_i\)) \(\to\) World at \(t_{i+j}\)
P3: No control about the Laws
P4: No control about the Past.
IC: Therefore, no control about the World at \(t_{i+j}\)
C: Therefore, no free will.

What to criticize?

Problem 1: Past and non-control

Let’s abbreviate “no one has, or ever had, any choice about whether \(P\)” as follows: \(N(P_0)\)

P1: \(P_0\) is a true proposition about a time in the past: 1 July 1900.
P2: No one has, or ever had, any choice about any past event because no one can change the past.
C: Therefore, \(N(P_0)\): No one has, or ever had, any choice about \(P_0\).

Problem 1 (continued)

P2: No one has, or ever had, any choice about any past event because no one can change the past.

P2 is irrelvant: Just because no one can change the past now does not mean that no one ever had any choice about it.

Problem 1 (illustrated)

flowchart LR
    S[Smith drinks]
    G[Gets in car]
    C[Crashes into Jones]
    T[Testifies he didn't have any choice about past because he can't control the past]
    S --> G --> C --> T

“It was dark and the roads were wet and I was driving super fast (and I was drunk) so not only (1) can I not control the past now but (2) I could not control what was happening while I was driving. Therefore, I never had a choice in the killing of Jones.”

Exercise

Criticize Smith’s argument.

Problem 2: Remoteness and non-control

P1: \(P_0\) is a true proposition about the state of the world 5 million years ago.
P2: No one can change events in the remote past because no human exists to change anything.
C: Therefore, \(N(P_0)\): No one has, or ever had, any choice about \(P_0\).

Fixes problem with previous argument: \(P_0\) refers to a time when “No one was around back then to have had such a choice.”

Incompatibilism

Incompatibilism states that there are no possible worlds where

the world is deterministic
human beings are free

Can we think of a world were this is false?

Yes

A world where humans beings always existed.
There would be no remote past
We would always have a “choice”

Incompatibilism revised

Incompatibilism states that there are no possible worlds where

the world is deterministic
human beings are free
there is a time \(t_0\) that occurs before humans exists

Compatibilism

Goals and agenda

Last time: incompatibilism (free will and determinism are inconsistent)
This time: compatibilism (free will and determinism are consistent)
Goals
1. Define compatibilism
2. Examine the arguments for compatibilism
3. Decide if we are compatibilists

Free will and determinism

Compatibilism defined

Compatibilism

the free will thesis and (2) determinism are consistent (both can be true)

At least one possible world where there are humanlike creatures who are free and the world is determinism.
It is possible that if the world is deterministic, then a person is free.

Compatibilism clarified

Does not say that our world is deterministic.
Does not say anything about if the world is indeterministic.

Compatibilism: Arguments

There are a few arguments for compatibilism

Paradigm Case Argument (PCA)
Argument from Conditional Analysis (ACA)

Paradigm Case Argument (PCA)

There is some debate about PCA is even saying.
It seems simple, but when criticized supporters of PCA say “that is not what I’m saying! You are strawmaning my argument!”
So, I’m not entirely sure what PCA is.

PCA

The argument has two steps:

Step 1: Argument about how words get their meaning to free will existing
Step 2: Argument that Compatibilism is true.

Step 1: Paradigm cases

There are cases (paradigm cases) where anyone who understands the meaning of a certain word would apply that word to those cases without any hesitation.

Suppose you see this object and I ask “what is it?”

Image of bird

You’d say “That’s a bird.”

Words and meaning

Many words obtain their meaning from (1) its repeated application to paradigm cases and (2) the teaching of the word to others by reference to those paradigm cases.

“Robin” gets its meaning because we repeatedly apply it to PCs and we teach you call this bird a “robin”.

Words and PCs

Therefore, the meaningfulness of a word, name, or other descriptive expression depends upon the existence of these paradigm cases.

If robins did not exist, then “robin” would lack meaning.

Free will is meaningful

“free will” is meaningful. It means “the ability of an entity to act in a way other than the way they acted.”

There is a difference between

Rejecting the free will thesis: “we are not free”, and
Saying the words “free will” have no meaning.

Free will exists

Since ``free will’’ is meaningful and words get their meaning from PCs, there must be PCs of humans acting freely (acting but they could have done otherwise)

Since there are PCs of free will, free will exists!

Step 2

Step 1 showed that free will exists (if it didn’t, the word wouldn’t have meaning).
Step 2 will show that compatibilism is true (I’m going to abbreviate this step)

Scientific investigation

Science can tell us if the world is deterministic or indeterministic.

Since we are free, science cannot show us that we are not free (even if the world is deterministic)
It can only show us what free will consists in

Example

Science can tell us that color is a perception of different wavelengths being refracted through our eyes, and stimulating photoreceptor cells in the back of the eye
It cannot tell us color does not exist.

Similarly, science can potentially tell us what free will consists in, but not that it does not exist (even if the world is deterministic).

PCA Summary

flowchart TD
  subgraph Step1
    M[Words have meaning]
    PC[Meaning from PCs]
    F[FW is meaningful]
    C["FW exists (PCs)"]
  end

  subgraph Step2
    S[D or I]
    DF[If D, then free]
    Com[Compatibilism is true.]
  end  
   M --> PC --> F --> C
   S --> DF --> Com
   Step1 --> Step2

Problem 1: Explosive

Just because a word has meaning does not entail that it has a referent (PC) or gets its meaning from that referent.
It would imply that every meaningful word refers to some real thing (e.g., “ghosts” ergo there are ghosts)

Exercise

Think of some examples of words that are meaningful but don’t refer to anything.

Problem 2: Compound words

Objection: Consider “golden” and “mountain”, these words may individually have PCs but what about “golden mountain”?
Response from supporter of PCA: compound words don’t have PCs, the simple words do!

Objection Again!

OK. But isn’t “free will” and “free action” a compound word?

PCA Summary

flowchart TD
  subgraph PCA
  subgraph Step1
    M[Words have meaning]
    PC[Meaning from PCs]
    F[FW is meaningful]
    C["FW exists (PCs)"]
  end
  subgraph Step2
    S[D or I]
    DF[If D, then free]
    Com[Compatibilism is true.]
  end  
  end

  subgraph Objections
    o1["O1: Explosive"]
    o2["O2: Compound words"]
  end  

   M --> PC --> F --> C
   S --> DF --> Com
   Step1 --> Step2 --> Objections

Argument 2: Conditional Analysis

Assumption: We act based on our desires

Let’s assume we act based on our desires.

flowchart LR
T[You]
D[Desire]
A[Action]
T --> D --> A

I desire X, so I try to do X.
Incompatibilist says our desires (and therefore our actions) are determined by LAWS + WORLD.

PT1: Free will defined

Point 1: We are free iff had we could have acted otherwise.

Perhaps an an uncontroversial definition of FW accepted by both incompatibilists and compatibilists.

PT2: Acting otherwise

If we are free iff we could have acted otherwise, what would make we could have acted otherwise true?

The compatibilist says this:

Point 2: We could have acted otherwise iff had we desired differently, we would have acted differently

So, they are saying:

I desired X, so I did X. But, if I desired to do Y, then I would have done Y.

PT3: Possible World Analysis

Point 2: We could have acted otherwise iff had we desired differently, we would have acted differently

How do we understand this part “had we desired differently, we would have acted differently”? (if I desired to do Y, then I would have done Y)

Whether we would act differently depends upon a counterfactual scenario (something counter to fact)
It depends upon a scenario where we have a desire different than the one we do in this world.

PT3: Possible World Analysis

To determine whether we would act differently (would have acted otherwise) if we had desired differently, we consider

the possible world we are in
a possible world with one small change (we have a different desire), and then consider
how we would act with that different desire

Example 1: Falling

Suppose the world (w1) is deterministic and you are falling from the sky and you desire to fall.

flowchart TD
subgraph w1
Y["You (falling)"]
D["I desire falling"]
A["Action: Falling"]
Y --> D --> A
end
subgraph w2
Y2["You (falling)"]
D2["I desire not falling"]
A2["Action: Falling"]
Y2 --> D2 --> A2
end
w1 --> w2

Even if you desired differently, you would not have acted differently (still falling) so you could not have acted otherwise so you are not free.

Compatibilist claim

But the compatibilist will say there are scenarios where we are free

We are free because we could have acted otherwise
We could have acted otherwise because if we had desired differently, we would have acted differently
And that is true because there are possible worlds (counterfactual scenarios) where our desires are different and so our actions are different.

Example 2: Choice

flowchart TD
subgraph World1
T[You]
S[Smurfberry]
BB[Booberry]
A[Action]
T --->|desire| S --> A
T ---> BB
end
subgraph World2
T2[You]
S2[Smurfberry]
BB2[Booberry]
A2[Action]
T2 ---> S2
T2 --->|desire| BB2 --> A2
end
World1 --> World2

I could have acted otherwise because if I wanted Booberry, I would have bought it.

Compatibilist insight

Even though

we always act based on our desires
these desires are determined by LAWS + WORLD
we are still free because
- we could have acted otherwise
- because if we desired differently, we would have acted differently

Free action

An action \(A\) from \(S\) is a free action if and only if had \(S\) desired to do \(B\) rather than \(A\), then \(S\) would have done \(B\) rather than \(A\).

Exercise

flowchart TD
subgraph World1
T[You]
S[Smurfberry]
BB[Booberry]
A[Action]
T --->|desire| S --> A
T ---> BB
end
subgraph World2
T2[You]
S2[Smurfberry]
BB2[Booberry]
A2[Action]
T2 ---> S2
T2 --->|desire| BB2 --> A2
end
World1 --> World2

Exercise

Create a diagram similar to the above (with a different action) that illustrates a case of compatibilist free action.

Common confusion

S could have done Y instead of X if there is a counterfactual scenario where S wanted to do Y and did Y.

When we consider whether someone would ACT differently, we consider whether their action would be different if they desired differently.

we don’t (1) consider whether there is ANY possible world we do what we want but
if whether there is a possible world where if we desired differently, we would have acted differently.

Conditional Analysis Argument (ACA)

P1: S is free if and only if S did X, then S could have done Y.
P2: S could have done Y if there is a counterfactual scenario in a deterministic world where S wanted to do Y and did Y.
P3: There are other counterfactual scenarios in deterministic worlds where S wanted to do Y and did Y.
C: Therefore, even if determinism is true, then S is free.

Problem 1: Two desires

Suppose in w1 we desire two cereals: BB and SB. But we pick SB

flowchart LR
T[You]
S[Smurfberry]
BB[Booberry]
T --->|w1| S
T --->|w1| BB

We are free to choose BB iff if we desired BB, we would have chosen BB.
But we did desire BB and we didn’t choose BB.
Therefore, we were not free to choose BB.

Response 1: Strongest desire revision

Rephrase definition of free will based on our strongest desire.

Free action

Assuming \(S\) did \(A\) rather than \(B\), \(S\) was free to do \(B\) rather than \(A\) if and only if had \(S\) more strongly desired to choose \(B\) over \(A\), then \(S\) would have chosen \(B\).

Assumes we always act based on our strongest desire
FW involves considering how we would act if we had a different strongest desire.

Strongest desire illustrated

flowchart TD
subgraph World1
T[You]
S[Smurfberry]
BB[Booberry]
A[Action]
T --->|strongest desire| S --> A
T --->|desire, w1| BB
end
subgraph World2
T2[You]
S2[Smurfberry]
BB2[Booberry]
A2[Action]
T2 --->|desire| S2
T2 --->|strongest desire| BB2 --> A2
end
World1 --> World2

Problem 2: Possible world talk

S could have done Y instead of X if there is a counterfactual scenario where S wanted to do Y and did Y.

The above is not what we mean (want) when we talk about free will
We mean “if we rewound time in this world, then could act differently.”

Compatibilist?

Exercise

Are you a compatibilist or an incompatibilist?