When we look over the history of English verse the steady progression away from the strongly rhythmical is marked, and yet there is, as far as is known, no explanation of why this should be so. It is obvious that modern poetry tends to be written in formless forms, by why has this happened, when the reverse movement, into ever more elaborate, regular, repetitive structures is, on the face of it, as likely. I suspect, in addition, that random movement is not the explanation, since the same direction of change is found in all the European languages, and also, perhaps in other literatures.

To return to the English case, let us observe that the change is heralded by a progressive abandonment of the ornate forms for serious statement, until in our period elaborated verse is perceived as risible, and if ostensibly serious, then ironized. This point begins to open up the field for us, since it appears that the transition is one of alterations in attributed value. At first we can do no more than say that the fashion for ornate verse has waned at the same time that its default perceptual state has become more and more humourous; but there seems every reason to think that these two facts are related. – Perhaps it is that standard verse has progressively been abandoned as a vehicle for serious thought because it has been adopted as a vehicle for comic expression. This is a little more plausible than the reverse position, i.e. that comic writers have merely adopted what was left open to them, but it still leaves unanswered the question as to why ornate writing should become “funny”. I propose, therefore, to treat these two phenomena, as some may think them, as one and provide only a single cause for both.

I shall begin with a discussion of what constitutes (elaboration) in verse, produce a logical consequence of this fact and then make some remarks on the possible historical significance of this point.

The ornaments of poetry are numerous, metre and rhyme, in all their many combinations, and tropes, in all theirs. For the time being I shall leave aside the question of tropes, which have largely survived the decline of metre and rhyme, and concentrate on these latter elements. Both metre and rhyme are repetitive structures. Not only do they repeat, but they do so at repeated intervals. This gives the reader the impression of a structure ordered on regular principles, a phenomenon that is not common in the texts which he is likely to read elsewhere, or to the utterances which he might here and there come across. Such things are rare, exhibit a peculiar degree of formal order and, because of their predictability, are oddly memorable when compared to other text strings or uttered statements.

We can appreciate just how unusual these forms are by a short train of reasoning involving the shelves of Borges’ much discussed universal library, in which there is a copy of all the books which have been written, all the ones which shall be written, and all those that could be written; in short, all possible books. (I am assuming that this is a grammatical library.)

Since the library is infinite we might expect there to be an infinite number of Miltonic sonnets in the library, and while it is true that there are an infinite number of books containing one or more sonnets in this form, the number of actual sonnets is finite. How can this be so? The infinitude of the number of books in the library is a product of two features, firstly that the length of books are not fixed, secondly that sentences may be infinitely extended by the addition of (framing) phrases, a phenomenon labelled 'recursion' by linguists and logicians. But if the extent of the text string is specified, or metred, recursion cannot operate, and the number of grammatical examples of English falling within that specification is finite.

A sonnet is a specified structure, having a fixed length of no more than 14 lines, each line to contain five beats (or just possibly, 6), and ten syllables (just possibly twelve). It will rhyme in a particular pattern, and though words can be multiplied infinitely in number, the line length limitation again means that there will be a limited number of rhymes.

Every additional specification makes the possible set of sonnets smaller, though of course, even with the numerous limits placed upon it in the various forms conventional among us, that set is still very large. The point is that it is finite, and that it is a member of an infinite set, the set of all grammatical sentences. Since this is so it is an infinitesimal: n/∞. Sonnets are very improbable forms, but so are any specified forms, pantoums, villanelles, triolets, double dactyls or whatever.

Now you may think that this is interesting as far as it goes, but has rather little bearing on the larger question of verse in general, since although there are a number of highly limited forms there are also an infinite number of possible stanza forms, rhyme schemes, and combinations of these parts. This is, of course, quite true, but the limitation issue applies at another level than that. I shall now argue that there are a limited number of poetic lines. In English it is possible to write lines having anywhere between one and six beats. A seven beat line decomposes into a four beat and a three beat line. These lines can employ either (duple) or triple metre, in rising or falling cadences. Practically, it is only possible to write three, four and five beat lines, this restriction being extrametrical and imposed by the lung capacity of human speakers or some other as yet undetermined factor. Given that there is this limitation on line length it therefore follows that there is a finite number of metrical lines, though there is of course an infinite number of combinations of these lines into poems of unspecified length. Just as with the sonnets this is also an infinitesimal.

It is now time to make a decisive move in the argument, and I will do so by suggesting that the limitations found in the forms honoured as poetic around the world all have the same root cause: they make the statement so couched and unusual in that they belong to a small set in the possible utterances available to speakers of that language. It is tempting to suggest that musicality is the main element in this matter, but there are reasons for thinking that this is not so. Firstly in some languages, such as Japanese, the verse tradition does not use restrictions concerning musicality, it counts syllables, and secondly musicality, euphony, and melody, need not be associated with metricated repeating structures, though in fact they often are. Indeed, some people would suggest that such repeating structures constitute musicality. I suggest that they do not, and that they are merely commonly associated, their appearance being explained through a similar line of reasoning given above.

I propose therefore that in all languages the principles of order associated with poetry will be found to create a finite set of possible text strings. The main claim rests on the assumption that the relative scarcity of strings which satisfy the conditions of the restriction makes them particularly interesting to the human mind, that is that these are items which the mind is particularly likely to respond positively to, and acquire. The question of memorability is a simple one; clearly texts with a repetitive structure, particularly rhymes, will have enhanced memorability, since certain elements of the information are encoded twice in the same form and related to each other. If one is lost the other, which points to the blank location, can be used to reconstruct at least some part of the information. It is noticeable that in really good mnemonics the rhymed words will be ones that are hard to infer from grammar and context.

However, the reasons that would explain why versified text is of especial interest, aside from memorability, are not evident to me. I have in other notes suggested that structural regularity is evidence of design, and the extremely designed appearance of verse may give it an enhanced truth value in the eyes and ears of the recipient, but I have no evidence on this point and cannot take it further at present. It may be that its “specialness” is all that is required to explain the fact, it may be that an insistent beat reduces resistance, or it may be that the listener mistakes the apparent external authority, by which I mean that the fact that the language permits a rhyme, may be seen to justify it.

Before moving on to the historical thesis concerning these matters, it will be necessary to make a preliminary logical move from the state of the argument so far. We have seen that the restrictions of verse are such as to reduce the set of available text strings from an infinity to an infinitesimal. We know, therefore, that it is easier to say something which does not conform to the restrictions than it is to say something which does so. – This much is obvious and commonsensical. I have discussed only a few of the restrictions used in verse, since I want the matter to remain clear, but there are, in fact, large numbers of supplementary restrictions over and above the simple metrical requirements already listed. Complying with these is additionally problematic. However, they are optional grace notes, and even very respected poets use them only occasionally in each line, or group of lines. I am thinking here of the (ingenious) variations in stress position and strength that prevent monotony but do not break expectations created by the metrical set.

Thus we see that statements in verse are variably satisfactory, that is they are variably elegant, or proper, in their satisfaction of the requirements of the specified repetitive structure. Colloquially we say that a piece of verse is more elegant, or better, than another.

Now, statements and groups of statements are also variable in their truth value, i.e. in the degree to which they program a brain to model a specified physical process with accuracy and success. (Alert readers will note that I am trying to side-step the objection that statements are only variable between two positions, true and false, and that I am doing so by suggesting that such exclusive truth analysis is only appropriate as a debugging technique for discussing the elements of propositions.)

The next stage is to ask whether these two variables are positively correlated. Our first response must be, of course, that they are not, and there is a mass of empirical evidence to suggest that they are not so. However, we can from earlier stages of my discussion show that they cannot be so. We know that the set of metrical lines is finite, and that the same is true for most specified lyric forms, and that even longer forms, epics even, may be rendered finite sets if the structure is so specified as to present the recursive possibilities of language infinitely expanding its length. All the members of these sets are more metrical than the non-members of any of these sets. Therefore if truth value and metricality were positively correlated all the members of these sets must have a higher truth value. Since, however they are a finite set and there is an infinite number of situations to model this is highly unlikely.

We are therefore confronted with two genuine variables, structural repetitiousness and truth value. Two variables cannot be simultaneously maximized. All possible statements can be placed on a variable truth axis, and they can also be placed on a variable structural repetitiousness axis. On the truth axis some statements in verse are higher in value than some in prose. But if we want to maximize the truth value of a statement we will have to neglect the structural regularity of that statement. Consequently if you wish to maximize the structural regularity of a statement, this will be at the cost of its truth value.

This point, which lies beneath all suspicions that verse, indeed any utterance which appears to be maximizing a variable other than truth, is trifling and a waste of time, is almost commonsense to those educated in other disciplines, but it barely occurs to those in English, or if it does it occurs as a guilty fear that they are wasting their own and their student’s time. However, I propose that this state of affairs is relatively recent and is the result of the growth, firstly in literacy , and secondly in that offshoot of widespread literacy, the scientific world and its information exchange. The growth of these two areas has made comparison between on the one hand, prose, the written and less redundancy laden, version of everyday speech, and, on the other poetry, both possible and necessary, and since the mnemonic value of repetitiously structured utterance is now negligible the judgement will often go against the poets.

I am suggesting that the conflict between prose and poetry, so evident in day to day life, even for a literary professor, is based on not on an arbitrary preference but on a correct estimate of its probable truth value. Moreover, there is a prevalent suspicion of those who employ such forms, and this too is soundly based. Though it is, as noted above, true that some statements in verse have a higher truth value than some statements in prose, it is also necessarily true that for every statement in verse there must be a statement in prose with a higher truth value, though lower memorability. Therefore whenever a person chooses to use verse and rhyme they are maximising some other than truth value, and though this will be in many cases acceptable, in comic verse for example, in others the audience will be well advised to stand on their guard.

You will now appreciate the point to which I am driving. The decline in the status of verse and rhyme can be explained as a decline in its utility, the powerful mnemonic effects being rendered insignificant by printing and literacy, and as a growing awareness that formal repetitiveness is inevitably correlated with reduced truth value. As this status declined poets, anxious to maintain their position, altered their writing techniques so as to retard the erosion of public respect, and writing styles which had previously been scanted because they were less ornately structured became more prominent. The history of English verse is the history of this movement, through from the exceptionally elaborate structures of mediaeval written verse, through the stanzaic elaborations of the 17th Century to the camouflaged repetitiousness of the 18th Century couplet, and so on into the 19th Century and the strong current, moving from Shakespeare, via Milton, into free verse. The triumph of blank verse and its high status, was rapidly and in some senses paradoxically achieved, for what could be more amusing in fact than the jingling bells of rhyme? It is barely surprising, however, that this form should be associated with a dramatic and wih an epic philosophico-religious poem, for in both cases the need confronting the poet will be to produce statements that do not appear egregiously false, and this is difficult since both employ extended narrative and we are, for good reasons, well-equipped to detect flaws in stories.

Detailed descriptions, with accompanying statistics, will have to wait, but a few closing remarks on the result of that history, in our own time, are worth making. Contemporary poetry is predominantly non-metrical, or so subordinates its principles of metrication that they are a minimal restriction. Yet the pretensions of its makers, and the claims of its admirers have not diminished, indeed they have increased in volume and intensity, as well they may, for no one listens. The poets are in a difficult position: they wish to be taken seriously and regarded as a distinct and distinguished statement producing group, but the most obvious way in which their product may be so distinguished is now, properly, recognised as a structural form that sacrifices truth value for memorability and effects which would perhaps have once appeared mystically profound and now seem merely comic. Verse and rhyme are the property of the humorist and the versifier, he who acknowledging the lowered truth value of his remarks employs these devices to entertain. What can the poet do? Firstly he or she may bury or camouflage the principles of metrication, thus hoping to produce a text string that will slip under the guard of the now-alerted reader. (I should say that I do not mean to suggest that poets are fully aware of the character of their actions, they probably perceive their actions as entirely aesthetic, by which we may understand that the first person who is to be deceived is themselves.)

Secondly, they may accentuate other rhetorical elements while reducing the repetitious structural complexity or removing it altogether. For example it would be unsurprising to find that the sentence structure of prose poems are not only more elaborate than that of ordinary prose but also that of verse. Other elements of style for which similar points might be made include, density of simile or metaphor (if this were born out by research it would be a remarkable triumph for my theory, for the richness of verse in these elements is proverbial). The use of less restrictive repetitive structures such as assonance or consonance instead of metre and rhyme, the use of obscure reference to subdue the reader, and similar postural devices which transfer the burden of distinctiveness from the work as it can be seen on the page to the writer. This latter idea is of considerable interest, since the theory seems able to explain what is at first sight an oddity, that as poems have grown simpler, in the obvious structural ways, the claims made for poets, have become grander and grander.

If I am right, these activities are aimed at distinguishing poetry, perhaps only subliminally, without alerting the reader to the fact that truth value is not high on the agenda. We should therefore expect to find apparent freedom from constraint, or partial freedom, such as blank verse and imperfect rhyming more commonly than complete absence of restriction, since these features, if smuggled through, may have the desired effect on the reader. Alternatively we may expect to find that elaborate tructures are used, but the poems are presented as heavily ironised verse. – I would call these pieces “Mock Comic”, to indicate that there is some doubt over the modesty and humour claimed in these poems and for the statements they make. A variant of this position may be to render ludicrous the idea of truth, thus insulating the poem against rebuttal.

A further possibility is that other principles of order will be employed, such as semantic metrication, the recurrence of lexical items or semantic fields at intervals more or less regular.

The possible substitutions are numerous, but as can be seen they are all either attempts to disguise structural repetitiousness and its constraints, to replace it with a principle which may not appear to interfere with truth value, or to abandon it entirely and transfer the major claims for importance from the text to the poet.

These are stop-gaps merely and cannot address the underlying question of truth value and utility, without, that is, abandoning “poetry” altogether and turning to prose. To do this of course would be to adopt the techniques of science. The alternative to this, already hinted at, is to abandon the high claims made for poetry, and to do so genuinely and not in the insincere “Mock Comic” and ironic mode already mentioned.

Instead poetry can be regarded as a form of statement not principally concerned with truth value, and that openly and for other, unpretentious purposes, maximises other variables. Suitable purposes would be humour and mnemonic value.