Sunday, October 27, 2013

Freedom of the Will and the Concept of a Person: Frankfurt

Introduction and Context
Up until now we've looked at a quick overview of the free will vs causal determinism debate.  Aside from the two basic positions--causal determinism and libertarian free will, there are two other "meta-positions": compatibilism and incompatibilism which are basically attitudes toward the debate.  

Incompatibilists say that causal determinism and free will are mutually exclusive:  only one can be true.  Compatibilists say that free will and causal determinism are not mutually exclusive:  Free will just means that the agent, rather than some external entity, was the causal origin of an action--For example, a free action can be free even if it was causally determined by the agent's genetics or atomic make-up, or whatever.  In other words, for compatibilists, a free action is one that wasn't caused by constraints or forces external to the agent.

We've also looked at a third position--Strawson's Basic Argument--in which he argues that we are not morally responsible for our actions regardless of whether determinism is true or not.

In this next piece Frankfurt argues for a different compatibilist position from the traditional one.  He says that mere absence of constraints or external causes isn't sufficient for free will.  What is needed for free will is for the agent to have desires about which desire motivates their action--this is what he calls second-order volitions.  

He takes the following approach to formulate his position:  (a) identify what distinguishes a "person" from an animal or non-person, (b) suggest that that quality is free will, (c) show why his conception of free will is superior to others, (d)  explain how his notion of free will relates to moral responsibility.

Freedom of the Will and the Concept of a Person

What Makes a Person Different from an Animal? 
The first step is to figure out what trait it is that we implicitly refer to when we identify someone as a person.  One way to do this is to compare the qualities we imply by "person" to those we imply by "animal."

Often we use the word "person" to mean the singular form of "people".  On this interpretation we are simply using the terms to distinguish members of  our own species from other species.  However, there is another way to use "person."  By "person" we also mean the quality or cluster of properties that distinguish us from animals.  For Frankfurt, this is, "the structure of a person's will."

What is "the structure of a person's will"? To figure this out, lets look at animals first.  Animals can act on their desires. Maybe they desire some sort of food or even have some higher level desire like the desire to be part of a group.  So, merely having desires can't be what distinguishes persons from animals.   What distinguishes our will from an animal will is the type of desires we can have; specifically, second-order desires.

Having second order desires means we can desire to desire things.  For example, maybe I'm not a very friendly person because I don't desire to be friendly.   On the other hand, I realize that this isn't a very good thing, so, I can want to want to be friendlier.  This type of second order desiring is pretty typical.  It often happens with food...

We can also desire not to desire things.  For example, maybe I like ice cream just a little bit too much.  I wish I didn't.  I can desire that I don't want ice cream so much.   ("I wish I didn't want ice cream so much" or "I wish I wanted salad more that hamburgers").  The bottom line about second-order desires is that we can reflect on our current set of desires, evaluate them, and form second-level desires about those desires.  Animals can't do this.  They only have 1st-level desires.  What makes a person different from an animal, in the sense that Frankfurt is talking about, is that a person can have second-order desires.

To better understand second-order desires about someone else's 1st-order desires, you should listen to this song:

The Will for 1st Order Desires
There are two ways to interpret desire statements (A wants X) as they relate to first-order desires.  In the first case when I say "I want to eat tofu" this can mean that I want to eat tofu but I also have a collection of various other desires about what I want to eat for dinner, and wanting to eat tofu is only one of them (maybe I want to ice cream and cake for dinner as well or instead of tofu) .  Maybe, the ice cream and cake win out.  It doesn't mean I didn't have the desire to eat tofu.  It just wasn't as strong as the other competing desires.  Basically, at any given time we can have a whole bunch of desires about what we want.  Some them may be mutually exclusive or some may be complimentary or some may have nothing to do with the others.

On the other hand, when I say "I want to eat one more dorito" this can mean that my desire to eat the dorito is the motivating force for my consequent action: it is evidence of my "will".  So, on this interpretation of "I desire X," the desire of X is the thing that causes me to act.  When someone asks me "why did you eat the dorito," I answer "because I wanted to."  The will is the first-order desire that "wins" the motivational competition with the other competing desires and causes you to act.  When people talk of the will, this is usually what they mean--but this is not what Frankfurt means by freedom of will.

Plain Second Order Desires Vs Second-Order Volitions
Frankfurt makes a distinction between two types of 2nd-order desires:  one that illustrates freedom of the will and one that doesn't.  In the first case we have a desire about which of our 1st-order desires will cause us to act.  In the second, we want to have the experience of having a particular 1st-order desire, but we don't want to actually act on that first order desire.  If you're confused like me the first time I read this, lets look at some examples to help make sense of this:

Lets look at the latter first.  To illustrate what Frankfurt means by a 2nd-order desire that doesn't identify what he means by free will, he uses the following example:  Suppose a doctor that works with heroin addicts believes that he'd be better able to help his patients if he knew what it was like to desire heroin.  Presumable, he'll be better able to help his patients if he understands what their addiction feels like.  He has a desire to have a first-order desire.

Now, just because he desires to know what it's like to want to take the drug, it doesn't follow that he actually wants (1st order) to take the drug.  He just wants to know what it's like to want heroin!   Taking heroin wouldn't satisfy the desire he has because his is only a 2nd order desire, not a first order desire (confused yet?).  He has no firs- order desire to take the illicit drug.  If you put heroin in front of him, he'd decline because he doesn't want heroin, he only wants to know what it's like to want heroin (to have the 1st-order desire).   He has no desire to actually be motivated to action by the 1st-order desire.

The previous account isn't the type of second-order desire that Frankfurt has in mind when he talks about the will.  Second-order desires that describes the will are ones where an agent desires to have a first order motivating desire that they either don't currently have or that isn't presently sufficiently strong to motivate action.  Consider a tofu example.  "I want to eat tofu" is in my (1st-order) set of desires concerning what I want to eat for dinner--but it's not the only one.  I also want to eat cake and ice cream.  The different desires don't all have the same strength.  Only the one that eventually moves me to act is my first-order motivating desire (i.e., my first-order will).

I can have a second-order desire about which of my dinner desires is the one that will eventually win (i.e., motivate me to act).  If it were up to me to program what 1st-order dinner desires I have and the relative strength of each, I'd make it so I really really want to eat tofu--much more than ice cream and cake.  I want for my desire to eat tofu for dinner to be the one that wins the motivating battle with the other desires.  

This is what Frankfurt refers to as a second-order volition (as opposed to a plain 2nd-order desire).  A second-order volition is a second-order desire that a particular first-order desire motivate you to act. A second-order volition is saying "I want this first-order desire to be the one that causes me to act rather than some other one."  Frankfurt also equates second-order volitions with the type of will that persons have (not to be confused with undeliberated 1st-order wills that animals also have).

Wanton vs Person:  The Essence of Person Isn't Reason but Will
Beginning with (and possibly earlier) the classic demarkation between persons and non-persons (i.e., animals) has been rationality.  Frankfurt wants to show that this criterion falls short.  To illustrate his point he distinguishes between wantons (someone with no 2nd-order volitions) and persons (someone with 2nd-order volitions).

A wanton is a type of dumpling commonly found in a number of Chinese cuisines.  They can be filled with different types of meat and can be steamed or cooked in soup.  They are often confused for persons which are also filled with meat but should not be steamed or cooked in soup.  Don't confuse the two.

Frankfurt makes a further distinction.  A wonton is someone who is controlled by their 1st-order desires.  They make no attempt to exercise their (second-order) will.  They may have second-order desires about their first-order desires but they have no second-order volitions.   They make no attempt to change what their first-order desires are or their relative strengths.  They don't take into account whether the 1st order desires are "desirable" or whether their relative motivating strengths are desirable.  

However, this is not to say that wantons are irrational agents.  They can reason about how they will carry out their desires and what will be the best way to do it; i.e., they have "practical reason".  They can also have competing 1st-order desires (I want to eat the chocolate bar vs I don't want to eat the chocolate bar) but they have no second-order desire about which of the two 1st order desires they want to win.  They do not reflect or deliberate on which 1st-order desire might be preferable.

Wantons can be distinguished from persons not by the capacity to reason but by the (lack of) exercise of second-order volition.  A person might have the same set of competing 1st-order desires as a wonton. The person both wants and doesn't want to eat the chocolate bar.  The difference between the person and the wanton is that the person wants the "don't-eat-the-chocolate-bar" desire to win (i.e., to motivate action).  She can and does reflect and deliberate on which 1st-order desire might be best to have.  The preferred 1st-order desire might not win in the end, but the fact that the person has the second order desire about which one wins is all that matters for being a person--not rationality which the wonton also posses.

Persons, Second-Order Volitions, and Freedom of Will
Consider the person (not the wonton) who has two conflicting first-order desires:  to eat the chocolate bar and go to the gym.  He has the second order-volition to have the go-to-the-gym desire.  If this desire motivates his behavior; that is, if he goes to the gym, then we can say that he was exercising his will and acting according to his will.

If he ends up eating the chocolate bar, then we say he acted against his own will.  What he really wanted was to be motivated to go to the gym.  However, for whatever reason--brain chemistry, conditioning--it didn't work out that way.  We can say, when he eats the chocolate bar that, "the force moving [the person] is a force other than his own, and it is not of his own free will but rather against his will that this force moves him to eat."

Freedom of Will Vs Absence of Constraints (Classical Compatibilism) 
Classical compatibilism defines free action as any agent's actions that were not coerced or constrained or caused by elements external to the agent.  Frankfurt says there's a distinction between freedom of will and acting free from constraints.  Consider that animals are free to do as they please.  Shoot, my dog still pees in my house if it's rainy or too cold outside--even though I'm sure he knows he's not supposed to.  But just because animals can pee wherever they please and not being coerced into where they pee, we don't typically say they have freedom of will.

To further bring out the distinction we observe that it's possible to prevent someone from doing whatever they please without undermining their freedom of will.  A prisoner may not be able to act in certain ways but they can still control which desires they do or do not want to have and which ones they do and do not want to try to act on.  

Freedom of will is when there is alignment of our second-order desire about our first-order desires with the first order desire that actually ends up motivating our actions.  In short, it's when the desire that we want to motivate our action ends up being the one that motivates our action.  A person can be free from external constraints to do what they want yet not have freedom of will if they fail to have any desires regarding which 1st-order desires motivate their action.  Conversely, someone could have freedom of will but not be free from external constraints.

"Complexities" that Can Lead to the Destruction of Persons 
If the whole desiring desires thing wasn't complex enough for you, here are a couple of possible extra-complex problems:

A.  Conflict of Second-Order Desires:  It could happen that I'm conflicted over which of my first order desires I want to motivate me to action.  In such a case, I'd have a conflict between second-order desires about my first-order desires.  If this happens, you cease to become a person because you have no clear second-order desire with which your first-order desires can conform.  You've lost your freedom of will.

B.  Infinite Regress Problem:  Suppose your second-order desires conflict, then you may have volitions or desires about which second-order desires you'd like to win.  But what if your tertiary-order desires also conflict about which second-order desires you'd have?  This could, in theory, go on infinitely. If this is the case, then your personhood is also destroyed because your freedom of will is lost.  There's no correspondence between the higher-order desires ('cuz you haven't decided on one yet) and the first-order desire that motivates action.

Partial Solution to B:  Just because the regress is possible doesn't mean it will necessarily happen.  It may happen.  But it may not.  Anytime a decision is made between competing higher-order desires about lower-order desires, the regress stops.

For example, suppose I am asked if I want to want to eat the chocolate bar.  If I indeed do want to want to eat it, then the regress stops.  The question "do you want to want to want to eat the chocolate bar?" has already been answered the moment I make a decision concerning the 2nd-order desire to want to eat the chocolate bar.  The regress only begins in cases where I don't or can't make a decision either way.

What Frankfurt's Theory Explains:  
Why Freedom of Will is Desirable:  Unlike many other theories of free will, Frankfurt's theory explains why freedom of the will is desirable.  Freedom of the will allows for the satisfaction of 2nd-order desires.   Its absence means those desires are frustrated.

Why We Don't Typically Ascribe Free Will to Animals:  Another thing Frankfurt's theory explains is why we don't typically ascribe free will to animals.  Other theories of free will describe as miraculous the fact that, even though we are physical systems, we are able to break the causal chain.  We can seemingly move our limbs at will.  But so can animals!  We don't consider the fact that animals can move their limbs at their leisure evidence of free will.

The Relationship between Free Will and Moral Responsibility
As we've seen, contra classical compatibilists, it doesn't seem to follow from the mere absence of external constraints that an agent has freedom of the will.  And if we say that freedom of the will is required for moral responsibility, it follows that absence of physical constraints on a agent isn't sufficient for moral responsibility.

It also doesn't seem to make sense to talk about free will and moral responsibility from the point of view of 1st-order desires.  We don't choose our own original set of 1st-order desires.  Some of us like eating ho-hos and sitting on the couch, some of us don't.  Some of us like to exercise, some of us don't.  It doesn't make sense to praise the person who goes to the gym because their pre-existing 1st-order desire to go to the gym is pre-set to be stronger than their first-order desire to sit on the couch eat ho-hos.  This person just happened to have lucked out.  

Moral praise or blame comes into the picture because as persons we are able to have second-order desires about which of our first-order desires "win".  Maybe my disposition is to sit on the couch and eat ho-hos rather than go to the gym.  But I have the capacity to deliberate on this fact and will that, despite this, my desire to go to the gym be the one that motivates me.  This capacity to reflect on, deliberate, and desire that our pre-existing 1st-order desires be otherwise or be ordered otherwise is what gives rise to moral responsibility. 

It might be causally determined that we desire what we desire;  that is, we might be causally determined to have the 2nd-order desires we have.  Frankfurt doesn't think this is a problem.  Freedom of the will is simply the exercise of 2nd-order desires.  Their causal origins don't matter.  If you have freedom of the will, you are also morally responsible for your actions.  

Can we really want to want things?   The wants you are able generate has a lot to do with your psychology.   Maybe you can have second-order desires about 1st-order desires that you already have, but maybe you can't have second-order desires about 1st-order desires you don't have.  Can you "will" a 1st-order desire into existence?  Maybe you can only shuffle what you have.


  1. Your explanations made this piece much more clear. Thanks for posting!

    1. Thanks for letting me know. Glad I could help.

  2. Thanks for posting! It helped me understand this paper. But I am a little bit suspicious of what you wrote about The Relationship between Free Will and Moral Responsibility. I can not sure that whether Frankfurt will agree with the sentence "This capacity to reflect on, deliberate, and desire that our pre-existing 1st-order desires be otherwise or be ordered otherwise is what gives rise to moral responsibility." Frankfurt wrote that "It is not true that a person is morally responsible for what he has done only if his will was free when he did it. He may morally responsible for having done it even though his will was not free at all." in this paper. And he suggest example of willing addict whoes will is not free. In this case although his will is not free he can be morally responsible.

    1. Thank you for your comment. I'll be reteaching this article in the summer and I'll have a closer look at it then in light of your helpful comment. I'll then make any appropriate revisions. I appreciate you taking the time to point this out.

    2. very well articulated thank you.

  3. Do you have any entrie talking about Frankfurts paper Alternate Possibilities and Moral Responsibility