Logic and Literary Argument

This web page addresses some of the most important questions about the use of logic in literary argument. Key terms in what follows include argument, logic, circular reasoning, intentional fallacy, biographical fallacy, syllogism, true, false, valid, invalid, fallacy, premise, enthymeme, evidentiary fallacy, hyprspecificity, and straw man..

Literary argument, like all sound argument, should not simply assert beliefs and recount facts but employ logical reasoning. The most common logical errors into which literary arguments often fall can be understood as varieties of the fallacies known as circular reasoning, intentional fallacy, and biographical fallacy.  Since these are very important, indeed, fundamental matters in literary criticism, I offer the following in the hope that it will help.

1) Consider this syllogism.
A) All men are mortal.
B) Socrates is a man.
C) Therefore, Socrates is mortal.
This syllogism is valid. Validity is a logical category.  Validity refers to logical structure.  Its opposite is fallaciousness, also known as invalidity.

C, the conclusion of this syllogism, is true.  Truth is quality of reality, not of logic.  The opposite of truth is falsity.

Validity and truth refer to two fundamentally different kinds of understanding.

In the syllogism above, C is true if A and B are true.  If, however, "Socrates" in fact is your pet cat, B ("Socrates is a man") is false.  In that case, C is still true but the syllogism is fallacious.  We might be able to discover that your cat is mortal by succeeding in killing it, but we cannot prove that it is mortal solely by the logic given here.  The first point, then, is that no matter how logical the structure of your argument may sound, if the premises are not sound, the argument is not sound.  This may sound obvious, but in practice it can be a problem because our arguments are couched in much less clear sentences than those of the first syllogism and because our chains of logic are much longer than that of the first syllogism.

2) Even syllogistic logic, which is the simplest kind, can get confusing quickly.  Consider this syllogism.
A) All men are mortal.
B) Socrates is a dog.
C) (Therefore,) Socrates is mortal.
While it is clear that C is true, we can see on the model of the cat syllogism in (1) that this dog syllogism is fallacious (= invalid); that is, it is invalid to assert the conclusion that Socrates the dog is mortal based on these premises.  Now it is also true that (B') All dogs are mortal.  We could have constructed a valid syllogism to demonstrate the mortality of Socrates the dog by deleting A and including B'.  But we didn't.  B' was omitted.

In writing real rather than sample arguments, why might we have omitted B'?  Two reasons come immediately to mind.  First, we may have made a mistake, either in our thinking or in writing down what we were thinking.  Mistakes can be caught and repaired.  That is one prime use for thoughtful revision.  Second, we may have presumed that our reader would know that the syllogism from (1) above is famous and also know in the context of our writing that Socrates is a dog and so this syllogism that is invalid yet has a true conclusion might have been intended to be funny.

When logic is reduced to stripped-down forms, it may seem improbable that one could omit a key premise. (A premise is a proposition on which an argument is based.) In practice, however, we are drawn to omit premises all the time.  Here is an example.
A) All men are mortal.
B) Achilles is a man.
C) Yet Achilles is immortal.
This may sound like a syllogism but it is not. At least two issues here are noteworthy. 

First, at least one premise, the premise in which we might have asserted that certain types of heroes are immortal, has been omitted.

A syllogism with a suppressed (or omitted) premise is called an enthymeme.  Most ordinary and literary argument is enthymemic.  We do not tend to say the following: A) People eat to assuage their hunger.  B) I am a person.  C) I am hungry.  D) Therefore I eat.  We tend to say the following: B) Whenever I'm hungry, D) I eat.  Obviously the second formulation is much more economical and hence, other things being equal, aesthetically, and therefore rhetorically, more effective than the first.  To gain this rhetorical power, we tend to speak and write that way.  But when a premise is suppressed, we are unlikely to examine its truth value.  So, in speaking and writing enthymemically, we may slip into using unexamined premises that in fact do not support our conclusions.  One should not, as a writer, tease out every step of a chain of logic on the page.  The result would bore the reader.  But as a thinker, one should examine every step of the chain of logic, testing to make sure the explicit premises are true and valid logic employed.

Second, the word "yet" is not a mere alternative for the word "therefore."  It implies that the logical move from B to C actually suppressed at least one premise.  In fact, many premises were omitted if we take this deduction to be valid.  Those premises would include assertions defining "man" and defining "hero" as a particular subset of man and defining "mortal" as a probable rather than inevitable condition.  The revision of our understanding of the meaning of mortality is the key revision implicit in the logic of "yet."

3) The language of our ordinary logic is rife with metaphors that often go unexamined.  The syllogism in (2) in which Socrates is a dog would be valid if a dog were a dog in one sense and a person in another.  The phrase "he's a real dog with women" suggests the possibility that we could understand B metaphorically.  If we do, then not only is C true but the syllogism is valid.  In fact, even when we are unaware of it, we use metaphors (and other figures of speech) all the time in spoken and written discourse, including argument.  "She eats like a bird, but the men can't get enough of her.  I guess it's her bean-pole figure."  There is logic in this passage, of course, something about the practice of eating little resulting in low weight, about men liking thin women, and therefore men liking this person.  But there is much more than logic here.  "Bird" (in Britain, and the related "chick" in the U.S.) has a slang meaning of female as sex object.  "Eats" sets up "can't get enough of her" to resonate with the constellation of eating/liking metaphors ("That baby is so cute I could eat her right up!" or "I'm hungry for you").  In this eating/liking context, the "bean-pole" is not only something skinny but in particular something skinny that supports the growth of beans, that is, the production of food, albeit skinny food.  While our "she" eats little, precisely by that act she makes herself into food, albeit skinny food again.  The speaker, in choosing a pejorative term ("bean-pole") rather than an approbative term (say, "willowy"), manifests disapproval of "she."  Implicitly, "she" is being called unnatural (a woman acting like a bird) and accused of being engaged in self-debasement, in self-objectification, done presumably to get the attention of men.  Whether or not our speaker is jealous as well as disapproving I'm not sure without more context, but jealousy on the speaker's part is clearly possible.  In short, the two-sentence argument conveys much more meaning of many different sorts than the first syllogism ever could.  Naturally we tend to speak and write in this richer way.  As this analysis suggests, it is precisely because rich writing sneaks meaning in below consciousness that we produce and consume it so enthusiastically.  Yet much of what such rich language says (for example, that it is wrong to objectify oneself for others) might warrant examination if only the argument were more explicit (for example, is it wrong to be a fashion model?).  The inevitable presence of rhetorical devices in our arguments opens the road to fallacy.

4) Circular reasoning is a common logical fallacy.  An example comes near the end of chapter 5 of Lewis Carroll's Alice in Wonderland in which Alice has a debate with a bird.  At this point, because of the fantastic workings of Wonderland, Alice has grown so tall and her neck so long that her head pops up in the tree tops.  The bird believes Alice is a serpent seeking the bird's eggs and tries to drive her away.  Alice refuses this rejection because, after all, she's not a serpent but a little girl.  The bird asserts that Alice is not little.  Well, not now Alice admits (but of course without meaning to surrender her own sense of identity as a little girl).  And Alice does go in tree tops, the bird declares.  Well, now I seem to, Alice admits, but not usually.  Well, if you're not a serpent, then you don't eat eggs.  Oh, yes, Alice says, I eat eggs.  Then, the bird concludes, you're long, in tree tops, and eat eggs, so you're a serpent.  What's going on here?  Carroll is demonstrating that the meanings of words vary from user to user and different viewpoints may have practical utility even without logical validity.  Obviously Alice and the bird have not agreed on the meanings of serpent or girl.  From the bird's viewpoint, Alice is a serpent because she has three characteristics that are necessary and sufficient to define a serpent.  From Alice's viewpoint, those characteristics are neither true of her (she isn't usually tall) nor sufficient to define a serpent (serpents, for example, don't have legs but Alice does, even if the serpent hasn't noticed them because the foliage obscures the bird's view of Alice below her neck).  As Alice sees the debate, the bird has asserted a definition of serpent (which Alice happens to reject) and then scolds Alice for fulfilling that definition.  Where did the bird get that definition of serpent?  She got it from her understanding of what matters in the world as she sees it.  In other words, the bird's conclusion that Alice is a serpent is based on premises that the bird wants to impose on Alice's understanding.  To put this more generally and in terms of literary argument, the bird reads the world as providing a certain definition of serpent and from that reading the bird concludes that Alice is a serpent.  The problem is, the bird's inference of the definition is wrong and her understanding of Alice is limited.  The bird's definition is, in fact, itself an assertion that requires logical support, but there is no logical support for the definition.  The definition is merely the bird's inadequate reading of the narrative world, her imposition on that world of her own understanding.  The definition comes first and then the supposed proof is based on that definition.  This is proving something (at the end) by making logical deductions from premises that themselves contain the conclusion.  Looping from the end to the beginning that way is called circular reasoning.  Circular reasoning often sounds right, as it does to the bird, but it is invalid nonetheless.  It is invalid not because the definition of serpent is false (which it is to humans) but because the structure of the reasoning is basically this: I see the following; therefore the following is what I am seeing.  That is circular reasoning.  It is often hard to recognize reasoning as circular because the steps between the first and last may be many or because we may have the feeling that there are suppressed premises that are nonetheless pleasant and clinch the connections.  But once we notice circular reasoning, we need to repair it because circular reasoning is always invalid.  Even if its conclusions, like the conclusion that Socrates the cat is mortal or that the bird sees what she considers to be a serpent, is true.

5) Literary argument often falls prey to a variety of circular reasoning called the intentional fallacy.  In the intentional fallacy, one infers from a text that the author "intended" to accomplish a certain goal and then uses that inference in interpreting other parts of the text.  In a famous instance, many people inferred that "The Shortest Way with the Dissenters" (1702) was simply too violent in its suggestions of how those who dissented from the authority of the Church of England should be treated.  Indeed, some clergymen said from their pulpits, that while dissent was obviously corrosive, such strict measures, although attractive, had to be understood as un-Christian and rejected.  Other clergymen, though, did accept these measures as necessary.  Only when people realized that the author was Daniel Defoe, a famous Dissenter, did those clergymen realize that the pamphlet was a satire.  They had interpreted each part on the ground that they already knew the author's intention.  Of course, what it turns out they knew was their own desire to see Dissenters as warranting stern treatment.  The misreading, no matter how psychologically understandable, was logically invalid.  It was an argument that reached conclusions (rejection or agreement with the proposals) based on an initial premise about the author's intention.  As a matter of logic, it is simply invalid to infer intention within a text and then on the basis of the inference interpret other aspects of the text so as to demonstrate that very inference of intention.

6) Literary argument often falls prey to evidentiary fallacy.  This is a term I use for mistakes based on the illogical application of extrinsic evidence.  The most common evidentiary fallacy involves authorial intention.  One might agree that the intentional fallacy is disabling because it draws inferences from intrinsic data to support conclusions about the very document in which those data are intrinsic.  But what about using extrinsic data?  What if an author writes a letter to a friend and says what was intended in the book?  Well, such extrinsic evidence is certainly suggestive.  A good critic will seek to see if that intention seems to have been realized in the text and, if so, to what effects, and if not, why not.  But while extrinsic evidence of authorial intent can and should be suggestive, we cannot infer that the extrinsic intention actually demonstrates that something in the text should be interpreted a certain way. This is so for three reasons.  First, the author could be lying.  Authors, after all, make things up for a living.  William Faulkner consistently asserted that the reason his characters with the same names changed from book to book is that the characters lived in his head and, like people, they just changed during the time between writings.  Had Faulkner shown any other signs of hearing voices, one might have taken this as in some sense true; however, at best one can call Faulkner's response an evasive metaphor.  At worst, Faulkner was just lying.  But he did it consistently.  We can't, in other words, simply trust an author to tell us what he or she is really trying to do.  Second, the author could have had the stated intention but failed to achieve it.  If authors could always fulfill their intentions, they would never write any book worse than their best.  Third, the author could be mistaken about his/her own intention.  In the "eats like a bird" passage above, it is easily conceivable that the person who we can see may be jealous of the slender "she" might have no conscious recognition of harboring that jealousy.  In short, while seeking knowledge of intention is normal and useful, even extrinsic knowledge about intention cannot be used to demonstrate logically that a certain bit of text must be interpreted a certain way.  It may, though, add weight to the suggestion that it is plausible to interpret the text a certain way.  Plausibility in this case, belief, may require a logical leap.  There is nothing wrong with asking one's own reader to make a logical leap sometimes.  We do that often.  I can't demonstrate syllogistically that it is right for me to support my political party.  If I could, I would lay out that argument and persuade everyone.  But I have made what I consider a reasonable leap of logic, inferring, among other matters, that I can more or less trust the leaders of the party to try to follow successfully their stated intentions which in some degree accord with my wishes.  I would support my party affiliation, should we debate about it, using logic among other tools, but I would not try to argue that logic alone requires my choice.  To make a deduction from an extrinsic statement of intention requires a leap of logic that claims the stated intention was both honest and fulfilled. This cannot be supported by the statement and logic alone, in part because if the supposed proof of its fulfillment is the text we are interpreting, that leap of logic constitutes circular reasoning. Thus, extrinsic statements of intention alone cannot demonstrate how one must understand an aspect of a text. Trying to do so is an evidentiary fallacy.

7) "The biographical fallacy" is an evidentiary fallacy. While intention can cause us logical problems either because of the intentional fallacy (which concerns intrinsic data and is a variety of circular reasoning) or because of an evidentiary fallacy (which concerns the limits to the application of extrinsic data in interpretation), the biographical fallacy is simply a variety of evidentiary fallacy.  It is common for people to say, for example, that Poe had to write about failed romantic relationships because his very young wife died very early in their marriage and he never remarried.  Well, if having only one important romantic relationship in one's life required that one write the almost hallucinatory story "Ligeia," large numbers of people would be doing just that.  But, one might add, Poe was also an alcoholic.  Perhaps, but by some estimates so are ten percent of the adults in America.  Where are their publications?  Biography, in other words, like extrinsic evidence of intention, can be very suggestive, often quite usefully so, but it can never definitively settle a logical argument about interpretation.

While it is common in teaching literary criticism to highlight the intentional and biographical fallacies, it is even more important to understand the nature of logical reasoning and the proper uses of evidence more generally.  Although powerful critics write powerful prose, both as readers and as writers, we should not allow the attractions of rhetoric to mask problems in logic.

8) Hyperspecificity, although not a formal logical fallacy, amounts to a logical fallacy in argument.  A hyperspecific statement is one that, while true, is so limited that in a sense it amounts to a false statement.  If you ask me what kind of man my father was and I reply only that “He had a moustache,” you might well think either that I didn’t like my father if that is all I can say about him or that I have some private understanding of what it means to have a moustache.  In either case, though, my reply did not constitute for you the questioner a reasonable response to the question.  Reasonableness, which has to do with a shared understanding of the world and how it works, is a looser concept than formal logic, but it is nonetheless compelling in argument.  Hyperspecificity, then, while not a formal fallacy, implies a lack of reasonableness in the larger communication situation in which an argument is made.

The straw man tactic in argument in some ways seems related to the use of hyperspecificity.  A straw man is a misrepresentation of a position in order to make that position easy to attack.  Consider the following dialogue (which I found 12 October 2007 at http://www.nizkor.org/features/fallacies/straw-man.html and to which I have added line numbers):

1. Bill and Jill are arguing about cleaning out their closets.
2. Jill: "We should clean out the closets. They are getting a bit messy."
3. Bill: "Why, we just went through those closets last year. Do we have to clean them out every day?"
4. Jill: "I never said anything about cleaning them out every day. You just want to keep all your junk forever, which is just ridiculous."

Line 1 defines the situation.  It implies many logical constraints on the speakers.  In line 2, Jill makes an argument (which happens to be an enthymeme).  In line 3, Bill, also suppressing many premises, implicitly characterizes the cleaning frequency Jill prefers (apparently at least once a year) as "every day."  This distortion is false; it was not her position.
In line 4, Jill implicitly characterizes the cleaning frequency Bill prefers (apparently no more than once a year) as never.  This distortion is also false; it was not his position.

Hyperspecificity may seem to be a straw man since it constitutes, within the communicative context, a distortion, but hyperspecific argument needs to be distinguished from straw man argument for two reasons.  First, a straw man is false while a hyperspecific statement is true.  My father really did have a moustache while Bill never said he wanted to keep his possessions forever.  A hyperspecific statement only amounts to a false statement; it is not actually false.

Second, hyperspecificity is a particularly seductive pitfall in literary argument.  Observing the vivid impact on the reader of the crimes Griffin commits in H. G. Wells’s The Invisible Man after Griffin has made himself invisible, and observing Griffin's stated goal to institute a “Reign of Terror,” it is easy to conclude that Griffin is an overreaching scientist whose science has alienated him from the community and thus opened the way for his immorality.  Since books are complex, we often may find evidence in them, among all that is within them, that we can use to support the hypotheses we have about them.  However, we need to test those hypotheses.  In this case, while the observations adduced are true and support the hypothesis, there is other evidence in the book that is relevant to this hypothesis, for example, Griffin being an albino which we learn caused others to alienate him in his youth and which is the final picture he presents in his broken-bodied death.  In other words, to call Griffin an overreaching scientist and stop one's characterization there misses the point that he was a person whom the community first alienated and who then, after years of this alienation, began his crimes.  Wells seems to place the blame on the process of alienation, which is shown to be self-perpetuating, and in this instance begins with the community, not with the scientist.  So, the failure to take into account enough of the relevant facts has led to a hyperspecific characterization of Griffin and thus has led to a false conclusion about the book.

One should note, in terms of formal logic, that while the conclusion is false, the argument is not strictly fallacious.  That is, in a syllogism, if one has a true major premise and a false minor premise, a false conclusion is valid (that is, not fallacious).  Given the stated premises, it is valid (correct according to the formal rules of logic) to draw the conclusion that egotism and science cause alienation in Wells’s novel.  But not to notice that in vivid ways egotism and science are not the first or only cause of alienation in the novel misses the main point and mischaracterizes the novel.  In Mary Shelley’s Frankenstein, egotism and science cause destructive alienation; in The Invisible Man, that position is smartly contextualized.  To fail to see the contrast between these works fails to understand Wells's position and achievement.  But that is not a failure of formal logic.  That is why I say that hyperspecificity amounts to a logical fallacy in literary criticism but is not actually a logical fallacy.  Nonetheless, it creates argumentative problems because it arises from a failure to test one’s hypothesis about a work against all the relevant aspects of that work.  Hyperspecificity, which appeals to us when trying to confirm a hypothesis, may lead us into error.  Hyperspecificity can be prevented best by seeking counterexamples, a process that has the additional virtue of deepening our understanding of the text.

Copyright © 2005, 2007 Eric S. Rabkin This page was last modified