2010/02/01

Is Law Based on Logical Fallacy?

A friend of mine, John Wolfgram, posed this interesting question:

It is true that argument ad verecundiam (from authority) is a logical fallacy and it is also true that argument from and to legal authority is the basic modus operandi of the law. Does that therefore mean that the basic method of operation of the law based in a logical fallacy?


The answer is not a simple "yes", because there is more involved than argument ad verecundiam. Legal practice is mostly about deciding what actions to take on behalf of the public, which involves what are called deontic propositions. During the course of that process, there is a tendency to indulge in making declarative propositions, assertions about what is or is not, and that is where such fallacies can be a special problem.

A court is a deliberative assembly of individuals with various duties and the authority to decide certain kinds of legal issues presented to it, called its jurisdiction. One of those duties is to preside over the court, and that presiding officer is typically called the bench or the "judge" (although it may be a panel of several, one is normally the chief).

One of the things courts are typically asked to do is "find" declarative propositions: that the defendant is or is not guilty, or at fault, or whatever. Obviously, the court can err, and declaring something true doesn't make it true. What the court is really doing is deciding "we will act as though it were true". They may have the authority to do that.

So when in law someone argues from authority one is really saying, "we don't know if what he is saying is valid, but we have to make a decision, and he seems more credible than the alternatives, so we will act as though what he is saying is valid." No ad verecundiam fallacy in that.

The fallacy comes when one transitions from "we will act as though what he is saying is valid" to "what he is saying is valid".

Let's examine what Court Y in Case B is doing when it cites a Court X in Case A as precedent. It is saying:

1. The evidence and arguments in Case A are similar to those in our Case B.

2. Court X in Case A decided to act as though the arguments for their decision were more valid than the arguments against.

3. We have confidence in the integrity and competence of Court X.

4. We don't have time to re-examine all the arguments in our Case B going back to first principles.

5. Therefore, in the interests of clearing our docket so we can take other cases, we will act as though the decision of Court X in Case A was valid, and act as though it is similar enough to our case B to make a decision to act in a similar way.


Now all of the above seems sensible, and indeed, given crowded dockets and limited cognitive capacity of judges, it would seem that "justice" could hardly be done otherwise if we are to have any finality in cases. However, as a matter of logic, every one of the five points above involves at least one logical fallacy, not only ad verecundiam.

So are we doomed to a judicial process that is logically infirm? Not quite. Game theory provides some useful insights.

Considered as a game, skillful play involves what are called heuristics -- decision strategies that do not guarantee always finding the best move, but which are highly likely to find an acceptably good move most of the time, and which are computationally tractable given constraints on time, cognitive capacity, and completeness and reliability of information about the state of the game.

Deontic logic, which is an extension of the first order predicate calculus, is designed to help us manage this kind of game playing, to optimize outcomes in general and over the long term. No guarantees of just decisions in every particular case, but a high likelihood of reasonably just decisions in most cases. The above five points represent attempts to use heuristics, whether with great skill and integrity or not.

Our job, in trying to decide and guide public policy, is therefore not to seek to impose strict first order logic on all of the decisionmaking of judges, but to improve the quality of such decisionmaking without excessively consuming scarce resources of dockets or personnel, and that involves reducing the susceptibility to error in each of the above five points. ("Error" being the euphemism for everything from incompetence to corruption.) By thus disaggregating the processes involved, we can better discern what might be done to improve them.

Now of course an important way is to improve the competence and integrity of judges. There is no substitute for that. But we also indeed to adjust our expectations of them.

Consider the recent case of Citizens United v. FEC. For me, with my background in constitutional study, it is an easy case. If I were to write the majority opinion in it, it would be very short:

1. The statute was passed by Congress.

2. The statute abridges freedom of speech and press.

3. The First Amendment says "Congress shall make no law .. abridging freedom of speech or press".

4. There are no later amendments to the Constitution that would supersede the First Amendment.

5. Therefore, the statute is unconstitutional.

6. Judgment for plaintiff.


It took me less than three minutes to write that. No need for extensive written or oral arguments (although I read the written arguments in about an hour). I can generally do the same with almost any case that turns on constitutional issues. (Statutory or regulatory construction can take longer.)

If we had nine justices on the Supreme Court with my background and skills, we could decide all 8000 cases submitted each year and the Supreme Court at least would not be a bottleneck. Now of course it would only be a handful of libertarian constitutionalists that would applaud those decisions. Most others would probably be very upset, if only because so many opinions would likely consist of only a few lines like the above. Not much for them to chew over, and their professional standing depends on having lengthy court opinions to chew over.

I am not unique. I'm sure we could find enough others like me to fill all the judicial positions. Of course, reliance interests, especially prosecutors, would go nuts.

But we can identify a large part of the problem. Judges want to avoid criticism, so they might overdo their deliberation and opinion writing. Now, most of the ones we have presently need to take even more time deliberating, because their subject matter knowledge and reasoning ability is weak. But that could be solved by requiring them to be much better educated. They should also be socially, as well as financially, isolated from the many interests and their lawyers that might appear before them, without lacking practical experience of many kinds.

13 comments:

argumentics said...

Nice insight. Still, something is fishy. Actually, a few things are fishy. First of all, ad verecundiam is not a logical fallacy; as you noticed in one of the examples, there's nothing "logically" wrong with a syllogism of the type: "Everything judge X decides about a problem in the field F is true. Judge X decided that Y is guilty. Y's guiltiness is in the field F. Therefore "Y is guilty" is true". To repeat, there's nothing logically wrong with such inference.

Second of all, the ad verecundiam is not always a fallacy. In fact, it is actually interesting that John Locke - from whom we have today this terminology - did not regarded the ad arguments as fallacies at all. In some contexts - that is, in some dialectical games - appeal to authority is a legitimate argumentative move. In most of the cases, it becomes illegitimate only if the authority invoked is out of the field (i.e. if proposition 2 in our syllogism is false and veiled).

And thirdly, argumentation in juridical context is not quite like argumentation in a critical (usual) discussion. It's more like a debate, where we have 2 parties and a deliberation. In other words, argumentation is aimed at the deliberator (judge) not at one party's antagonist(s). So fallaciousness has little space to occur. We can talk about false evidence, about proof, about claims and their warrants, about rebutals - but we can hardly talk about fallacies (seen as dialectical shifts in an argumentative discussion). I'm not that sure about this, but I know irrelevant claims are dismissed in some courts (Canada, maybe? I'm not sure). So informal fallacies (i.e. relevance problems) would not have any ground from which they could appear and rise.

Jon said...

No, argument from authority is still a logical fallacy, no matter how credible the authority might be, or how persuasive the argument. That doesn't mean it may not be properly done as a way to improve the odds of making a good decision, as a heuristic. But the basis for that is not logical inference, but probabilistic weighting.

The syllogism argumentics offers is not the one that is actually used. What is used is one of the form, "Judge X holds Y and Judge X is knowledgeable and skillful in field F and the holding is in field F, therefore the holding is true." In other words, the first premise is more like "Judge X is probably right." No one really asserts that "Everything judge X decides about a problem in field F is true." Any reasonable person acknowledges that in principle judge X could be wrong, even if no one can find an instance in which he has been wrong so far. Only if the authority is infallible in principle could the syllogism offered by argumentics be used.

argumentics said...

Jon, I must say you have a very interesting way of reading into concepts such as "logical" and "fallacious(ness)".

So tell me, what is illogical about "probabilistic weighting" (i'm guessing you meant induction). Induction is not to be counted as a "logical inference"?

Of course I constructed my syllogism as a dummy to play with. Since judges are not infallible instances, my example fails to be relevant. But still, there's nothing logically wrong with my example; if we accept all its premises, we must accept that the conclusion follows necessarily (again, not that this helps us in our discussion about decisions in court). Regardless, it is nothing logically wrong with that syllogism (it is a valid deduction). If you thing otherwise, please let me know which logical rule does it break.

So I agree with your restatement of our syllogism when discussing the special context of courts. Let's rephrase the syllogism in your way:

"Judge X holds Y and Judge X is knowledgeable and skillful in field F and the holding is in field F, therefore the holding is probably the right one"

This is an appeal to authority? Is this still a fallacy? Why?

Jon said...

No, inductive reasoning, while it may be persuasive, is not logical inference, in the sense of either the first-order predicate calculus (FOPC) or deontic extensions of it.

What is being done in inductive reasoning is the introduction of an inferentially independent model which seems to explain, and perhaps also predict and enable control, the data leading up to it. Model-building and confirmation or refutation of models is not logical inference as that is meant in deduction.

Deductive logic may be persuasive, but not all arguments that are persuasive are logical in that sense (and the only sense a logician respects).

There is, however, still a formal problem with your "syllogism" as you posit it, and that comes with the fuzziness of the sets involved, and the interpretation of the expression "probably the right one", which presumably translates into a statement like "the probability of decision Z being correct is > 0.5" and the interpretation of "judge X is knowledgeable and skillful in field F" as "the probability of correctness of most decisions by judge X in field F is > 0.5", which is really a composition of three statements, "the probability of the decision being in field F is > 0.5", "the probability the decision was made by judge X is > 0.5", and "if the decision is in field F and made by judge X the probability that it is correct is > 0.5". (And we haven't even gotten to how one decides whether a decision is "correct", which is another probabilistic statement.)

In other words, one has to be very careful to disaggregate all the probabilities involved and how they compose. A chain of probabilistic statements may compose multiplicatively, and the product of probabilities > 0.5 may be < 0.5. Without assigning numbers to them, the "syllogism" becomes nonsense at the logical level.

We see this among lawyers who seem fond of what I call the "analogic syllogism": "if A is similar to B and B is similar to C then A is similar to C". That might work for the restricted meaning of "similar" in plane geometry, but doesn't work (is a fallacy) for the broader meaning it takes in law. (In formal terms we say that the relation "similar" is not transitive across the full range of ordinary meanings of it.)

If we want to get even deeper into this, we would need to distinguish probabilistic statements from level of confidence weightings, which have the form of probabilities but are not probabilities in the strict sense as we derive it from measure theory, which requires operational tests like trials of large samples. That is where the derivation of probability theory from information theory can be a better approach, because information theory does support confidence weights in a way that measure theory does not. But that would take this discussion into a realm where we would need to be able to use notation this forum doesn't support.

argumentics said...

What you're saying is that induction is not a logical inference because it fails to comply with fopc, modal logic (of any kind) and that model building is not logical in the deductive sense. Right?

But this is because inductive logic is not fopc, modal, or "as is meant in deduction"! Of course it isn't. I didn't say induction is deduction or any other species of logic.

Nevertheless, induction is a type of inference, which makes it a type of logic (again, not the deductive, let's call it "analytical" type).

Jon said...

Some, like argumentics, try to extend the meaning of "inference" to include "inductive inference" or "probability logic", and as I have indicated, this can be done if done skilfully, but I prefer not to use the term "inference" for anything but the deductive kind, and to use the the term "inductive reasoning" but not "inductive inference". To use the term "inference" in that broader sense is to try to invest such reasoning with a rigor it doesn't have, which can be, and usually is, disingenuous.

I once took the LSAT exam used to qualify prospective law students, and found, in the "logical reasoning" sections several questions, that when cast into the rigorous forms of the FOPC, either had more than one, or no, correct answers. When I complained about that to the authors of the exams, they replied that the test was only of "informal reasoning". To this I replied that such "informal reasoning" permits contradictions, and thus allows any proposition to be both true and false at the same time.

That gets to the concern of Wolfgram that began this topic. If law is not based, at some level, on at least deontic logic, then law is not the application of rules but the whims of political judges, and is a huge fraud on the people, who expect more of it.

One can see this kind of "informal reasoning" used to "interpret" the Constitution. A strict adherence to historical evidence and deontic logic would not permit much of today's constitutional jurisprudence, and most of the abuses now being committed by government officials and agents. Yet by the kind of reasoning argumentics seems to be defending, one can "infer" everything those officials and agents are doing to be constitutionally authorized.

I don't think so, and I don't think most of the American people do, either, even if most of them are not skilled in deductive reasoning. They demand a legal order in which some legal propositions are valid, and some invalid, and that the two can be distinguished by reason and evidence alone and not on who is closest to the judge.

argumentics said...

I must place all my confidence in the last part of your comment. My readings in the area are pretty sparse.

However, since we dwelt upon the subject of fallacies, here's a pretty example:

"To use the term "inference" in that broader sense is to try to invest such reasoning with a rigor it doesn't have, which can be, and usually is, disingenuous"

My answer is "no, to use the term "inference" in that broader sense is to try to ... use the term "inference" in that broader sense and that's all". This quest for "underlying, unexpressed-but-I'm-sure-they're-in-there motives" is not something I presume you would give green light to.

Anyway, I enjoyed this conversation.

Steve Grensky said...

Hi,
I loved this post because I'm a federal lawyer and have posted comments VERY similar to your analysis of the recent FEC case in the context of health care on my blog steverg.blogspot.com
I NEVER read an opinion such as your proposed opinion even though yours is EXACTLY correct.
My analysis of the Constitutionality of national health care is:
1. Congress has only the enumerated powers in the Constitution pursuant to the terms of the document in general and, specifically, pursuant to the 10th amendment.
2. Congressional powers are listed in Article I, sec. 8.
3. National health care is not listed in Article I, sec. 8.
4. No amendment to the Constitution lists health care as a Congressional power.
5. National health care legislation is unconstitutional.
We can do this with almost any topic. It's very easy and the only logical way to proceed.
Let's take abortion.
1. Congress has only the enumerated powers in the Constitution pursuant to the terms of the document in general and, specifically, pursuant to the 10th amendment.
2. Congressional powers are listed in Article I, sec. 8.
3. Abortion is not listed in Article I, sec. 8.
4. No amendment to the Constitution lists abortion legislation as a Congressional power.
5. Any legislation attempting to govern abortion is unconstitutional.
The list goes on and on. With that above template, you can solve many of today's so-called 'difficult' issues.
The problem seems to be that people don't like the result so, rather than follow the amendment rules outlined in the Constitution, they just ignore the plain meaning and start dreaming up loopholes.
It's so refreshing to read a rational post.
Thanks for sharing!

Jon said...

In reply to Steve Grensky:

Thanks, but not all delegated powers are in Art. I Sec. 8. There are others scattered throughout the Constitution and properly ratified amendments.

As for abortion, Roe v. Wade was about state legislation, not federal. I discuss that in this blog article.

Steve Grensky said...

Hi, again

Yes, there are other Congressional powers scattered around the document, as you say, but it bogs down the analysis to add that a thorough search of the entire document reveals nothing.

Even your succinct 6 point analysis of the FEC case fails to mention the fact that other parts of the Constitution don't allow regulation of speech.

And, as for abortion, Roe v. Wade was about state statutes but I didn't mention Roe. I was only pointing out that if Congress tries to legislate on the issue, it is not within Congressional power to do so since it is not in the specifically enumerated powers found throughout the document.

If one wants to make it a federal issue, one would need to amend the document to do so. Stretching 14th amendment language or other references to 'persons' or 'people' expands rather than contracts federal power and, thus, should be avoided.

Jon said...

More broadly, the Constitution does not authorize the regulation of services generally, and medical services more specifically, on state territory. (Perhaps on the territory of federal enclaves as per Art. I Sec. 8 cl. 17.) This takes us into construction of the Commerce and Necessary and Proper clauses, about which see this article.

Kidgrok said...

Though a mere and humble member of the public whose welfare is addressed herein, I also love this post. I have no formal training in law but have come face-to-face with exactly the problem raised in Jon's treatment.

In particular, it is both infuriating and depressing to deal with an appellate court influenced more by public opinion than by case fact or Constitutional law. The hope for justice in that venue has turned out to be unfounded for the reasons expounded in Jon's analysis. Despite more realistic expectations in the submission of petition for review to state supreme court, I remain highly disappointed that there is so little adherence to the purer, more impartial path to justice. In review of their affirmative decisions I see so much arrogance toward the public from appellate--and even supreme--courts everywhere. The club to which all magistrates apparently belong is exclusive and lofty and shrouded under a veil of infallibility.

Moreover, subjective, morally biased opinions--often doled out by underlings and assistants in court systems--appear to weigh more heavily than rational logic in deciding an accused's fate. Evidence branded valid by a lower court's opinions, based more upon assumption than factual analysis, becomes "fact" thereafter and is never again questioned--or even considered questionable. This is poor policy when such evidence is patently misrepresented by a prosecutor in pursuit of charges against a defendant. Unless the probabilistic nature of both trial evidence and legal decisions is understood, acknowledged, and addressed, the validity of any guilty verdict will be far from reliable.

I applaud the time and effort Jon has expended in raising this question and presenting his answer. I only wish his erudite treatment could find many attentive and influential ears.

avamcitizen said...

I am curious if any of you are lawyers. These arguments about logic seem to ignore some time honored principles of the law..and the way it fuction in society. For example, the concept of stare decisis - meaning let the decision stand - is the idea that once a higher court in a jurisdiction has made a decision on a subject..that decision stands or binds all the lower courts..where the subject matter is the same. There is a very logical reason for this concept which I will not go into here. Sometimes courts will refer to the work of courts out of jurisdiction - simply because they think the primary issues are the same or similar and they agree with the outcomes. Out of jurisdiciton decisions are not binding on a court..they are just informative or persuasive. Finally, there are a variety of reasons why judges write long detailed opinions...some of them good..and some of them not so good. As far as the U.S. Supreme Court...I seldom find they're opinions to be overly superflous when you follow the threads of time and change..you see why. Although...they do make the wrong decisions sometimes. And I belive..in those instances it is because of outside influences..like politics...getting in the way of a reasoned and logical application of the law and precedent. This is especially true with Constitutional Law.

Translate

Follow by Email

Search this and affiliated sites

Blog Archive