Ermanno Bencivenga’s last book is Understanding Edgar Allan Poe: They Who Dream by Day, published by Cambridge Scholars. We offer here an excerpt of the book: its fourth chapter.1placeholder
A Sorry State of Affairs
Poe’s tactic in dealing with the questions I posed at the end of the last chapter [how true can we take a fictitious story to be? how can it increase our knowledge?] is two-pronged. On the one hand, he disputes the mainstream opinion that science, as ordinarily carried out, does unproblematically increase our knowledge. On the other, he provides his own alternative view of sound cognitive practice, for which we may or may not keep using the label “science.” In this chapter I will take up the first prong, and that will compel us to roam far and wide.
The contemporary understanding of how science works is the result of an important correction to Aristotle’s. Predictably, the father of logic saw a scientific discipline as logically textured: not as a wild bunch of true statements but as a hierarchical arrangement where conclusions follow from premises. This process of deduction, however, cannot go on in perpetuity, issuing in an infinite regress from some premises to their premises to their premises…; there must be a beginning to it, a top floor to the edifice. One must eventually arrive at principles which cannot themselves be deduced, which one must see to be true as they are evident (the word “evident” comes from the Latin “video,” “to see”). Therefore a good scientist, besides being proficient in logic, must have the virtue of noûs, which lets her see the principles.
A generation after Aristotle, the Egyptian mathematician Euclid (writing in Greek, the lingua franca of the Middle East after Alexander’s conquest) gave the most monumental application of this model in his Elements (of geometry). He started out with ten principles—five axioms, or common notions, of general scope, such as “If equals be added to equals, the wholes are equal” and “The whole is greater than the part,” and five specifically geometric postulates—plus a number of definitions; and went on to prove by logical argument 465 propositions, or theorems. His tour de force was to remain a matchless example of rigor for two thousand years: Baruch Spinoza’s masterpiece, written in the late seventeenth century (at a time when the Aristotelian system had already come under fire), was entitled Ethica, ordine geometrico demonstrata (Ethics, Demonstrated in Geometrical Order) and was structured to mimic Euclid’s Elements, with the entire apparatus of axioms, postulates, definitions, and theorems.
There was, however, a bit of trouble in paradise; there had been all along. Euclid’s fifth postulate, which in one of several equivalent versions reads “Through a point external to a straight line S goes one and only one parallel line to S” (and is thus typically known as the parallel postulate), did not seem to many as evident as the others; so repeated attempts were made over the centuries, by Greeks, Arabs, Persians, Italians, and other nationals to prove it from the sibling four postulates, thus reducing it to the status of a theorem and lowering the burden on noûs. One of the most extensive such attempts was due to Giovanni Girolamo Saccheri, a Jesuit priest and a professor of philosophy, theology, and mathematics at the University of Pavia, who published Euclides ab omni naevo vindicatus (Euclid Freed of Every Flaw—“naevus” in Latin is mole, blemish) in 1733 (the year of his death), supposedly deducing contradictions from denials of the fifth postulate. Since his deductions were faulty, what he had actually achieved was to develop a substantial amount of non-Euclidean geometry—geometry based on principles different from Euclid’s. But Saccheri’s work was not rediscovered until much later, hence it did not have any influence on the people who self-consciously performed that development: the Russian Nikolai Lobachevsky, the Hungarian János Bolyai, and the German Bernhard Riemann (Lobachevsky’s and Bolyai’s results were published well within Poe’s life but he never shows any awareness of them; Riemann’s appeared shortly after his death).
Philosophers like David Hume had already raised serious doubts that any empirical principle could be regarded as evident, and universally accepted as such. (Hume famously maintained that one could not be certain the sun would rise the next day: that we just have a habit of seeing it rise, and from that psychological condition nothing follows about the world.) Now the very Euclidean cornerstone of the Aristotelian conception of science had crumbled, and “principles” were found to be shaky in mathematics, too: no appeal to noûs could determine which of them were true. In 1899 David Hilbert, in his Foundations of Geometry, took stock of the situation, systematized the field by offering a rich smorgasbord of “principles” from which to assemble our plates (this combination will give you geometry A, that one will give you geometry B, …), and proposed that we think of a collection of principles (axioms, in his language) as the arbitrary rules of a game, which is up to any one of us to decide whether to play or not.
What is the state of the art today? In mathematics, Gödel’s theorems have established that, besides being optional in the sense advanced by Hilbert, axioms—and the axiomatic systems built on them—will never be secure: we will never know that the next thing we deduce is not going to be a contradiction. We can play whatever game we like, and abide by its rules, but we must be aware that the game may blow up in our face. (As Gottlob Frege’s did, when he tried to prove that arithmetic could be reduced to logic and ended up facing Russell’s paradox—which had been sitting all along at the core of his system.) In the empirical sciences, scientists and philosophers agree that axioms are to be judged on the basis of their consequences: if what is deduced from the axioms matches observations in the laboratory, we correspondingly get more confident in their truth. But, again, we can never be sure, since the next observation might cause a catastrophe. The hierarchical arrangement of a scientific discipline, its logical texture, is still there—this much of Aristotle goes still unchallenged. Without the naïve reliance on noûs, however, it is permanently, and irremediably, exposed to the risk of collapsing from the top. All in all, a rather sorry state of affairs, if one is nostalgic of apodeictic certainties.
Don’t take my word for it. Though scientists, especially when they act in some official capacity, have a tendency to discreetly pussyfoot around these issues, some of them are ready to call a spade a spade. Here is how Richard Feynman introduces a class of students at the California Institute of Technology to the “scientific method” in 1964 (the following year he would win the Nobel Prize for physics):
“Now I am going to discuss how we would look for a new law. In general we look for a new law by the following process. First we guess it. [The students laugh and he tells them: ‘Don’t laugh. That’s the truth.’] Then we compute the consequences of the guess to see what, if this is right, if this law that we guessed is right, to see what it would imply, and then we compare those computation results to nature, or, we say, compare it with experiment or experience, compare it directly with observation, to see if it works. If it disagrees with experiment it’s wrong. In that simple statement is the key to science. […] Now you see, of course, that with this method we can disprove any definite theory. If we have a definite theory, a real guess, from which we can really compute consequences which could be compared to experiment, then in principle we can get rid of any theory. We can always prove any definite theory wrong; notice however that we never prove it right. Suppose that you invent a good guess, calculate the consequences, and discover every time that the consequences you calculate agree with experiment. The theory is then right? No, it is simply not proved wrong. Because in the future there could be a wider range of experiments, you compute a wider range of consequences, and you may discover then that the thing is wrong. That is why laws like Newton’s laws for the motion of planets lasted such a long time. He guessed the law of gravitation, calculated all the kinds of consequences for the solar system and so on, compared them to experiment—and it took several hundred years before the slight error of the motion of Mercury was developed. During all that time the theory had been failed to be proved wrong, and could be taken to be temporarily right. But it can never be proved right, because tomorrow’s experiment may succeed in proving what you thought was right wrong. We never are right, we can only be sure we are wrong.”2placeholder
Philosophers of science have been no less peremptory—far from it. Karl Raimund Popper asserted that we can never verify a scientific theory (show it to be true) but we can falsify it (show it to be false)—if at least one observation impugns it. Thomas Kuhn described the “progress” of science as constituted of unmotivated, revolutionary jumps from one paradigm—that is, one vision of the world—to another, incommensurable one, where all criteria of correctness are internal to a paradigm and hence no paradigm can be judged except on its own terms (hence also one cannot really judge that from one to the other there has been any progress). Paul Feyerabend advertised anarchy as the best policy in the sciences and said that “the experience on which Galileo wants to base the Copernican view [defended in his masterpiece Dialogue concerning the Two Chief World Systems] is nothing but the result of his own fertile imagination, that it has been invented.”3placeholder
Sounds familiar? Of course it does. We saw [in a previous chapter] Poe declare that Kepler guessed his laws; and, because of his basic agreement on this matter with distinguished contemporary physicists and philosophers, he deserves his declaration to be given more context.
“Kepler admitted that these laws he guessed—these laws whose investigation disclosed to the greatest of British astronomers [viz. Newton] that principle [of gravitation], the basis of all (existing) physical principle […]. Yes! these vital laws Kepler guessed—that is to say, he imagined them” (E130).
In 1848 Poe wrote Mellonta Tauta (Greek for These Things Are in the Future), set a thousand years later in 2848 (though an initial note signed by Poe himself says that the manuscript from which it is drawn was found “about a year ago, tightly corked up in a jug floating in the Mare Tenebrarum,” F1117—as if the future it portrays was a metaphor of conclusions he had already reached, but that were, for all intents and purposes, still covered in darkness). In it, we read:
“It appears that long, long ago, in the night of Time, there lived a Turkish philosopher (or Hindoo possibly) called Aries Tottle. This person introduced, or at all events propagated what was called the deductive or a priori mode of investigation. He started with what he maintained to be axioms or ‘self-evident truths,’ and thence proceeded ‘logically’ to results. His greatest disciples were one Neuclid and one Cant” (F1120).
Aside from the bit of harmless fun Poe has with the corruption of ancient (and not-so-ancient) names (“the best names are wretchedly corrupted in two or three thousand years,” E124), and from his lack of recognition of Kant’s (or Cant’s) Copernican revolution (which was strongly indebted to Hume and let the Aristotelian system go), his picture of Aristotle (and Euclid) is fair. As is his criticism of this picture, when he writes (elsewhere) that “the Aristotelians erected their castles upon a basis far less reliable than air; for no such things as axioms ever existed or can possibly exist at all” (E126). Nothing, that is, is self-evident in the way the Aristotelians alleged: everything is open to disagreement, to a clash of beliefs.
“An axiom in any particular science other than Logic is […] merely a proposition announcing certain concrete relations which seem to be too obvious for dispute […]. Now, it is clear, not only that what is obvious to one mind may not be obvious to another, but that what is obvious to one mind at one epoch, may be anything but obvious, at another epoch, to the same mind. It is clear, moreover, that what, to-day, is obvious even to the majority of mankind, or to the majority of the best intellects of mankind, may to-morrow be, to either majority, more or less obvious, or in no respect obvious at all. It is seen, then, that the axiomatic principle itself is susceptible of variation, and of course that axioms are susceptible of similar change. Being mutable, the ‘truths’ which grow out of them are necessarily mutable too; or, in other words, are never to be positively depended on as truths at all—since Truth and Immutability are one” (E161-162).
Getting to particular instances of this general statement, the first part of the series The Literati of New York City (published in six parts, in 1846, in Godey’s Lady’s Book) opens with a discussion of Reverend George Bush, Professor of Hebrew at a local university, and of his work Anastasis, or the Doctrine of the Resurrection (1845), “in which it is shown that the Doctrine of the Resurrection of the Body is not sanctioned by Reason or Revelation” (E510). And this is Poe’s twofold critical judgment of the work:
“The ‘Anastasis’ is lucidly, succintly, vigorously and logically written, and proves, in my opinion, everything that it attempts—provided we admit the imaginary axioms from which it starts; and this is as much as can be well said of any theological disquisition under the sun” (ibid.).
In the second part of the same series critical scrutiny turns to another reverend, George B. Cheever, whose latest work mentioned is a Defence of Capital Punishment (1846). About which Poe has the same qualified attitude as about Bush’s Anastasis:
“This ‘Defence’ is at many points well reasoned, and as a clear resumé of all that has been already said on its own side of the question, may be considered as commendable. Its premises, however, (as well as those of all reasoners pro or con on this vexed topic,) are admitted only very partially by the world at large—a fact of which the author affects to be ignorant” (E529).
And, if there are no axioms—no absolutely firm points of departure for a demonstrative path—there are no demonstrations either: nothing that can show the truth of its conclusion (that is what “demonstrare” means in Latin: to show). There are inferences, no doubt, deductions, and they can be as tight as we want; but one thing they cannot prove is their very premises; and, as long as those premises are of indeterminate cognitive value, so is what we deduce from them.
“That the demonstration does not prove the hypothesis, according to the common understanding of the word ‘proof,’ I admit, of course. To show that certain existing results—that certain established facts—may be, even mathematically, accounted for by the assumption of a certain hypothesis, is by no means to establish the hypothesis itself” (E175-176).
“[W]hatever the mathematicians may assert, there is, in this world at least, no such thing as demonstration” (E122).
In the Republic, Plato speaks highly of mathematical entities. They are the closest thing to the forms: eternal and immutable as the forms are. So their study is the best introduction to the study of the forms, and future rulers must delve into mathematics for ten years before turning to the dialectic that will let them access the forms. But why, one might ask, are they just the closest to the forms, and why is their study just the best introduction to the study of the latter? What is missing from the study of mathematical entities? Plato’s instructive response is:
“[Mathematical studies] cannot yield anything clearer than a dream-like [!] vision of the real so long as they leave the assumptions they employ unquestioned and can give no account of them. If your premiss is something you do not really know and your conclusion and the intermediate steps are a tissue of things you do not really know, your reasoning may be consistent with itself, but how can it ever amount to knowledge?”4placeholder
At the origin of our philosophical tradition, in its inaugural text, the problem we have been debating, and Poe tackles with great energy and eloquence, was already present, and phrased in the clearest possible terms. But there is also another reason to bring up Plato here.
In the Theaetetus, while conversing with a young mathematician, Socrates is looking for a definition of knowledge. The best he (or, rather, his interlocutor) can come up with is “justified true belief”: for me to be able to say that I know a proposition A, A must be something that I hold to be true and is true, and there must be a good reason why I hold it to be true—not because, say, I flipped a coin and committed myself in advance to believing whatever the outcome of the flip was associated with. In an article of 1963, the American philosopher Edmund Gettier constructed some counterexamples to argue that justified true belief is not enough: mental experiments in which someone has justified true belief of A but we would not say that he knows A. So a fourth condition would seem to be needed for knowledge. A whole industry has grown out of Gettier’s, and similar, counterexamples, and untold candidates for a fourth condition have been proposed, generating no consensus. But we do not have to go there: the three conditions posed by Plato, which no participant in this debate contests, will suffice for my purpose. For that purpose is simply to ask, on behalf of Poe (and Plato): of the propositions of such a scientific theory as is described by Feynman (or Popper), can we claim we have justified true belief?
We might believe them, if we accept the theory as a description of reality. And we might be justified in believing them, if the theory has (as, say, quantum mechanics does) a high amount of empirical confirmation. But are they true? Maybe, and maybe not. So maybe they add to our knowledge, and maybe not. What is more: if they are not true, and do not add to our knowledge, we might eventually find out; if they are true, and do add to our knowledge, we never will. A sorry state of affairs indeed.
Under these dicey circumstances, it is time to revisit the scholars’ worry we are discussing here, and wonder: what does the presumed radical difference between science and literature reduce to? How isn’t a scientific theory but a tale we tell ourselves about parts of the world, imagining them to be made of things that look like elastic balls, or like ocean waves, or like solar systems where planets revolve around a star—and keeping faith with those images until, well, we just don’t? Or, to give the questions a positive spin, since we are about to turn to the positive side of Poe’s views, what would it take for a tale, told by Galileo or by Poe, by Newton or by Calvino, to graduate to a genuine contribution to knowledge? Fictitious they are all: they are all made, created (“facio” is Latin for “to make”) by their authors. But what are the criteria by which one such creation would be singled out as true?
Works Cited
E — Edgar Allan Poe, Complete Essays, Literary Studies, Criticism, Cryptography & Autography, Translations, Letters and Other Non-Fiction Works. An e-book: e-artnow, 2017.
F — Tam Mossman (ed.), The Unabridged Edgar Allan Poe. Philadelphia: Running Press, 1983.
Paul Feyerabend, Against Method. London: Verso, 1978.
In what follows, F abbreviates The Unabridged Edgar Allan Poe, edited by Tam Mossman. Philadelphia: Running Press, 1983, and E abbreviates Complete Essays, Literary Studies, Criticism, Cryptography & Autography, Translations, Letters and Other Non-Fiction Works. An e-book: e-artnow, 2017. At times, I have slightly modified a quote, upon comparison with other sources.
This lecture can be watched at www.youtube.com/watch?v=0KmimDq4cSU.
Against Method (London: Verso, 1978), 81.
Translated by Francis MacDonald Cornford’s (London: Oxford University Press, 1945), 253-254.