B. Lukács






Matter Evolution Subcommittee of the Geonomy Scientific Committee of HAS


H-1525 Bp. 114. Pf. 49. Budapest, Hungary





Part of the Sequence "Methodical Aristotle Studies"

Motto: The History of Science Is Not Science But Scholarship



            Diogenes Laertius recorded the total number of lines in Aristotle’s books. It is 445,270. Can we believe in this number? The number of extant lines is slightly above 110,000. If so, ¾ of Aristotle’s works is now lost. How it became lost, why and when?



            Most of Aristotle's writings are now lost. That is a shame: Antiquity was not careful enough abouts its greatest thinker, while practically everything is extant from Plato, inventor of the idea of myriad-year-old winged souls fleeing to above Sky, or the 9,000 year old Atlantean superpower defeated by heroic Athenians in 9,400 BC. Well, Fiction is generally sold better than Science, so it was copied much more diligently.

            My aim here is double: how much of Aristotle is lost, how and when they were lost. Chapter 1 sums up the size of the extant Aristotelian texts, Chapter 2 tries to guess the size in Late Antiquity. Then the ancient numbers are tested by Number Theory in Chapter 3. Chapter 4 gives some technical arguments about ancient data carriers, and the subsequent 2 Chapters try to reconstruct the process of data loss. The last Chapter is an Outlook: what could be done if anybody wanted to follow this line of thinking.



            Of course we know the extant works. However the size of them can be expressed in various units.

            The contemporary Aristotelica is published in 2 printed Volumes [1], on almost 2500 pages. That book is almost devoid of non-Aristotelian texts as Notes, Comments & such. There is cca. 44 lines on a page and cca. 78 letters in a line, so it is cca. 3400 letters on a page. So it is cca 110,000 printed lines or cca. 8,500,000 letters, including empty spaces between words & such. As for the Corpus Aristotelicum (CA; later), it is shorter. The texts of Aristotle fill 2463 pages, of which the CA is 2338; 95 %.

            Of course, this is the English text. The Greek can be slightly shorter, but the difference cannot be too much. However for manuscripts numbers can differ a lot. Now, in manuscripts the lengths varied. However in the ms. from which Bekker worked in XIXth century [2], the CA takes 1462 pages, and a page generally 2*38 lines (2 columns). That is almost exactly 111,000 lines (10 double lines being empty on the last page, and a few between Books), almost the same number of lines as in the printed English version. (Printed English lines have more characters.)

            The extant material is roughly from 3 sources. First, there is the Corpus Aristotelicum. As far as we know, this goes back to cca. 60 BC, and was edited first by Tyrannion, or by Andronicus of Rhodes, or both. The oldest ms. available now comes cca. from Xth century, and we may more or less believe that the Bekker canon represents CA. Modern scholars have various views about books how much they are "genuine". Indeed, some books may have been written by followers of Aristotle in the Peripatetic school, and surely most books were penned by students or young teachers. However that is trivial, as far as the book has been based on Aristotle's lectures. The books of the CA are as follows (+ means that it is probably "genuine", - that it is probably not genuine):








De Interpretatione



Prior Analytics



Posterior Analytics






Sophistical Refutations






On the Heavens



On Generation and Corruption






On the Universe



On the Soul



Sense and Sensibilia



On Memory



On Sleep



On Divination in Sleep



On Length and Shortness of Life



On Youth, Old Age,...



On Breath



History of Animals



Parts of Animals



Movement of Animals



Progression of Animals



Generation of Animals



On Colours



On Things Heard






On Plants



On Marvellous Things Heard









On Indivisible Lines



The Situations and Names of Winds



On Melissus, Xenophanes and Gorgias






Nicomachian Ethics



Magna Moralia



Eudemian Ethics



On Virtues and Vices












Rhetoric to Alexander






            This is 44 books, all of them very probably attributed to Aristotle even in Late Antiquity. As told, CA is some 95 % of the texts.

            The second source is a papyrus from Egyptian sand, found in 1891. The almost complete manuscript is known for us as Constitution of Athens. Of course it cannot have the canonical Bekker numbering, but it is some 2 % of the whole extant material. However, as we shall see, it originally was not a book on his own right

            The third source is Fragments; generally short texts of a few sentences extant in the texts of other authors. In addition, in any times there were texts attributed to Aristotle, as e.g. his correspondence with Alexander III of Macedon, or with Hephaestion. Such writings were quite popular in Middle Ages, when methods were taken, allegedly from the great Stagirite, about production of gold, or the Stone of Philosophers, or stories about Candace, the Indian Princess. I ignore such writings.

            So, again, the total amount of the extant works is cca. 110,000 lines, either printed or in manuscript.



            Lists of Aristotelian titles are extant from 3 serious sources. Diogenes Laertius gives a list of 156 titles organized into cca. 400 ancient books. Among them we can find titles similar or identical to the following members of the CA: 1, 2, 3, 4, 5, 6, 7, 11, 12, 14, 20, 21, 22, 27, 28, 30, 31, 32, 34, 36, 38, 40, 41, 42 & 44, albeit in many cases we may conjecture the different edition. Then comes the list of Vita Menagiana, with 9, 10, 24, 26, 29 & 35, together with unknown other books. (Vita Menagiana is probably the work of Hesychius of Miletus, only fragments are extant. Instead of looking for Muller’s Fragmenta, now you can consult with [1].) Finally, Life of Ptolemy [1] gives the titles of 8, 13, 15, 17 & 37.

            Now, ##16, 18 & 19 are missing, but they may hide behind other titles; the edition of Parva Naturalia looks rather arbitrary. #23 can be behind other Animalia books. #25 may be (or not) in On Vision, and #33 in Storm Signs. I am unable to find in the lists ## 39 & 43, but maybe my imagination is weak. This is the reason I told that Late Antiquity may have known all CA.

            As for the Constitution of Athens, Diogenes Laertius mentions a title: Constitutions of 158 States. From fragments we know some scanty statements concerning the following states: Acarnania, Acragas, Aegina, Carthago, Colophon, Corcyra, Elis, Gela, Ithaca, Locris, Massalia, Methonia, Naxos, Opuntia, Sparta, Syracusa, Tenedos & Thessaly. Obviously Constitution of Athens was also a part of it.

            For the Fragments, Diogenes Laertius knows about a book Protrepicus, and now there are even some attempts of reconstructing it.

            A lot of Diogenes Laertius' other titles are quite reasonable too. E.g. he mentions a "Claims", anything that be; and look, in Fragment 614 of Rose's compilation [4] Ammonius mentions Aristotle's Claims of the Cities in connection of Alexander of Molossus. In a lot of cases such independent mentions happen. So we may accept that the books mentioned by Diogenes & others indeed existed and now lost; it is another matter if Aristotle wrote them, or his followers or strangers.

            Now, from the 3 lists mentioned above, we can collect 211 titles, albeit a few of them may be variants. We now have 45 books. So some 1/4 of the books are extant and 3/4 lost. Very probably the more important books had higher chance to be copied, of course.

            Now, Diogenes Laertius happens to give the total number of the lines in Aristotle's Works: 445,270 lines [3]. Although we do not know the lengths of these lines, and they were certainly variable, the ratio of lost to extant lines is again almost exactly 3:1.



            It would be interesting to know how Diogenes Laertius counted the lines. Since it is irrealistic to assume that he read everything (plus he gives such numbers for a few other authors too), there remain two possibilities. Either some librarians (e.g. of the Alexandria Library) calculated the numbers from their catalogues, or he simply invented the numbers for making his book more interesting.

            His numbers are as follows [3]:



Lines total










c. 30,000


Now we can see that Crantor was not very productive (or his works were not carefully collected, hence the circa), Speusippus' low production is quite conform with my expectation, Xenocrates & Theophrastus were much better, and Aristotle published as much as these two together.

            Quite reasonable, but the problem is just that they are reasonable, so Diogenes Laertius may have invented them. And observe that the total number of Xenocrates' lines is a prime number!

            So let us see the primes in these numbers. They go as



Line numbers factorised into primes




224239 (!)








The last line is a variant; an Internet site [5] gives the total number of the lines of Theophrastus as 232,800 instead of 232,908. The other numbers are the same.

            Now, the first impression is that not only Xenocrates' big prime is improbable, but Aristotle's big 6361 as well. By other words, somebody invented some numbers for some philosophical purpose of his own.

            But beware! It is an interesting problem, which prime constituent is probable and which is not. Here I cannot solve this problem in full. However I can make random number calculations. I generated 10 6-digit numbers by means of generating each digit via independent random numbers from 0 to 9. The next Table is the 10 numbers and the prime decompositions, in chronological order:


Random numbers

Prime factors




















201389 (!)


            Obviously the sample is too small; e.g. from 10 numbers 3 start with 0, instead of the expected 1. However if someone needs a bigger simulation, it means only a moderate computer work. Taking these 10 numbers, one is a prime, quite comparable to Xenocrates' line number, and another one has a prime factor almost twice as big that the biggest in Aristotle's line number.

            So for me Diogenes Laertius' statement for the total lines of Aristotle does not seem invented. We may accept that it was originally 445,270, and then 3/4 of them is now lost.



            We have stories how the famous Alexandria Library bought Aristotle's books. Indeed, there was a time when the Alexandria Library collected anything. Also, the Peripatetics had some strong connections with Alexandria. Demetrius of Phalerum, the first Director of the Library, was in his youth the pupil of the Lyceum, and Theophrastus, second head of the Lyceum suggested him to Ptolemy I. And Strato, third Scholarch, was at a time the tutor of the Ptolemy Heir Apparent. So it would have been surprising indeed if the Library of Alexandria had not collected books of Aristotle.

            Indeed, the simplest assumption about the list in Diogenes Laertius is the Catalogue of the Alexandria Library. We know that real, topical and alphabetical, cataloguing was developed there. Antique sources tell us that the Catalogue in itself took 120 papyrus rolls. If Diogenes had access to the Catalogue, he could easily compile his list, with the total number of lines. If he had access.

            Now let us make again some calculations. You can get some useful information from [6], [7] & [8]. First let us note that in the antiquity Aristotle's books were scrolls, and in Alexandria surely papyrus scrolls. In the Pergamum Library some kind of pergamen was also possible. Romans used codices too, organised essentially as our paper books, but 99 % of extant Greek manuscripts from 2nd century AD are on scrolls [7]. To 5th century codices became predominant, but then we are well after Diogenes Laertius.

            Now, how big is a scroll? There is no definite answer; scroll sixes vary. However we know lots of "books" from ancients; small units of texts rather similar to our Chapters. The old texts were maybe copied and recopied many times, but they are still so organized, into these “books”. Take, for example, Aristotle's Problems. That is 38 “books”, going from 859a1 to 967b25, so almost 8200 lines; so one “book” there is only cca. 215 lines. Metaphysics is 14 “books”, from 980a23 to 1093b27, so cca. 8700 lines, so 620 lines per “book”. On Indivisible Lines is a single book, with 380 lines. And so on.

            Also, take Iliad. It consists of 24 “books”. Now, we have the line numbers as follows:





















































            Ref. [7] tells that "A thousand or so lines of text was all that a roll could hold...". Indeed, the above Table shows this. There are shorter Books, but the longest s between 900 and 1000 lines.

            However after some times Alexandrian scrolls became longer, surely because of the need of the Royal Library. Ref. [6] tells that a scroll could hold "one of the rather longer books of Thucydides". Now, the average book of Thucydides would be cca. 3500 lines in Aristotelian format. According to extant papyri with Greek texts 1000 lines needed some 7 m length (halved, of course, with double columns, but then a wide papyrus is needed), so a book of Thucydides would be 20 m, rather at the edge of technical possibilities (think about abrasions in evolution/involution).

            Now we can re-check Diogenes Laertius' line number 445,270. Assume that the Alexandria Library had an edition "Complete Works of Aristotle", as Ref. [1] now, only bigger four times and on scrolls. Then of course they may have copied all Aristotelian texts onto "standard" scrolls. Although the scrolls could not be uniform in modern sense, they may have had a "nominal" length, and then Diogenes could have produced the total line number by simply multiplying the number of scrolls with the capacity of a scroll. However from prime factorisation the only possibility would be 70 scrolls, each with 6361 lines. But that would mean 40 m long scrolls, seriously improbable. (And even this improbable explanation would not work for Xenocrates.) It is much more probable that the 445,270 is really a result of some ancient proto-statistician, from catalogues. Of course I cannot tell if the summation had been correct.

            We can, however, estimate the length of the part of the Catalogue about Aristotle, if Diogenes' data come thence. Minimally a few lines are needed for an item: author, title, topics, number of scrolls involved, place. If so, I would guess as 600 lines, a small older scroll. This means that Aristotle's books were well below 1 % of the Library. Of course, they were the most valuable books.

            Historians of Science often simplify the question about losing Aristotle's books. There was a big Aristotle collection in the Alexandria Library; it perished with the Library. Maybe there was another collection in the Pergamum Library; but Mark Anthony sacked that Library to give a present to Cleopatra. The collection of Lyceum vanished with the end of Paganism in 4th century. Western collections were burnt by Goths, Huns or Vandals. And so on. However the thing is more complicated. You can find a discussion of mine at [9], and I do not repeat that. However observe that

            1) The later history of the Alexandria Library is obscure (see later).

            2) In the historical folklore Mark Anthony sacked the Pergamum Library, but sent the books just to Alexandria.

            3) One may doubt if followers of the Peripatetics were still Pagan in 393. As for Logics, for example, in High Middle Ages Abaelardus (11th c.) and Petrus Hispanus (12th c.) stand on good Peripatetic grounds [10], [11], [12]. One can loosely span a bridge through Aristotle - Theophrastus - Poseidonius - Alexandrus Aphrodisius - Boethius - Simplicius - Philoponus - Abaelardus without anything drastic at the religious change; the logic remains on Peripatetic foundations [13].

            Of course Christian Peripatetics may have not been interested in Theses on Love (4 books), Homeric Problems (6 books) or Olympic Victors (1 book), so they did not keep them for us. And a copy of a scroll book deteriorated in some centuries, and the few copies of a rare book were endangered if libraries simply did not get founds.

            I shall suggest a two-phase process of losing 330,000 Aristotelian lines in non-copying. But first let us see the Fate of the Great Library of Alexandria.



            We are sure that sometimes there was a Great Library in Alexandria. It seems almost sure that it had been founded by Ptolemy I sometimes at the end of 3rd century BC or his son Ptolemy II Philadelphus, the first Director was Demetrius of Phalerum, the Kings were fanatic to develop the Library, once it had hudreds of thousands (!) of scrolls; and we (more or less) know the Directors until Aristarchus of Samothrace, the Grammarian (mid-2nd c. BC).

            So far so good. But the sources are equivocal about the number of scrolls. Ref. [6] compares some of them: the smallest number is given by Seneca (AD 50), 40,000, the largest by Aulus Gellius (AD 150), 700,000. We can conclude that data are simply hearsay and fairy tale for later authors.

            A moderately early source is Aristeas. This is the source describing the translation of the Pentateuch to Greek by 72 wise men, so arguing for the prestige of the Septuagint, which is now the original for the Catholic Old Testament (by no means the same as the Protestant Old Testament, or Jewish Tora, or Karaite and Samaritan Bible; the canon of Septuagint is the oldest). Aristeas tells that in a time the Library had 200,000 scrolls but later it went up to 500,000.

            I would suggest the lower value, since I am familiar with state-owned economies. There are plans; then local leaders do not fulfil the plans of the top, but they report that they have. Ptolemy II Philadelphos gave money and promised more, to get a very, very big Library. Then librarians sure rounded upwards all possible data, and reported their plans as facts.

            We heard about the Catalogue in 120 (large) scrolls. That would be cca. 420,000 lines. It is rather hard to catalogise 500,000 scrolls in 420,000 lines. It is barely possible for 200,000 scrolls. Maybe that meant cca. 50,000 artifacts, and, say 40,000 titles (duplicates must have existed).

            Now, these 200,000 scrolls needed re-copying from time to time. Under Egyptian climate papyrus is quite durable (the Oxyrhynchus excavations result in many quite legible pieces after almost 2 millenia), but the process of reading is quite dangerous for a long scroll. When fiscal troubles started for the Library, some books started to deteriorate.

            Romantic historians of Science are not interested in such details. They look for a single Catastrophe. The suggested events are various.

            1) During the military campaign of Caesar in 47 BC fires started and the books of the Library burnt.

            2) Emperor Caracalla punished Alexandria in 215 AD by massacre & looting for some imaginary sin half a millenium old. During the looting the Library was destroyed.

            3) The city revolted against Emperor Aurelian in 273 AD, so it was looted, the palace quarter became ruins, and so also the Library.

            4) Emperor Theodosius abolished the pagan temples in 393. Then the Serapeum (or the Museum) was overtaken by Christians, and they destroyed the Library.

            5) In 415 the mob, excited by monks of Bishop Cyril, lynched the pagan astronomress Hypatia, and destroyed her favourite place, the Library.

            6) The Arabs took Alexandria in 641. They did not consider the Library necessary, because the wisdom of Allah is simple, so they used the scrolls to heat the baths; they burned for 6 months.

            7) The crusaders burned the Library.

            Now, lots can be told against any of the above Solutions. For the first 6, however, at least the bare possibility remains. However, Strabo (true, ex silentio, but that is something for him, usually loquacious) is a witness against any Great Library in 13 BC [14]. In XVII.1.8 he simply states that

            1) A part of the royal quarter is the Museion.

            2) The Museion consists of a garden, an exedra, and a big residential place for the scientists staying in the Museion; the scientists eat there.

            3) The Museion has its incomes, and is led by a priest.

            4) This leader originally was nominated by the [Ptolemy] Kings and now [13 BC] by Caesar [the Emperor].

            Indeed, no mention of any Great Library, although it would be proper because of the scientists. The mention of the scientists, however, is important: the Musaeon of Alexandria continues as Research Center under the Romans.

            Obviously the scientists need a Library. Now, there are calculations for the size of the Great Library; mainly from the fact that the books must be housed along the walls in a convenient way. Ref. [6] suggests a method of calculation, which gives 20 m*20 m*5 m for 200,000 scrolls. Such a substantial building should have been recognised by Strabo, visiting Alexandria as a friend of the Roman Legate.

            Strabo's silence is interesting; one would expect him telling more about the Research Center. But clearly, his text is against a Great Library of the above size. But: he tells where the scientists eat, but not, where they read and work. Strange enough. Scientists of course need hotel & cafeteria; but also offices & library.

            But surely, no Great Library in 13 BC. Then either Caesar's Alexandrian War burnt it in 47 BC, or it vanished even earlier. Now, remember that we know the names of the Library Directors until mid-2nd century BC, and they were scholars or scientists. The last "great" librarian is Aristarchus of Samothrace [6]. He seems to have been the librarian to the time of Ptolemy VII Physcon; the time of this King was turbulent, with counter-Kings, Syrian incursions, crises between Alexandria & Chora. There are some stories how scholars were molested & expelled. Surely, the Library went on; we know about the existence of Ptolemy VII Physcon’s officer & librarian Kydas, and about an Onasander in 88 BC. However no scholarly or scientific tradition remembers them; so in the best case they were administrators. In the worst case they stole the money of the Great Library; or they were honest, but the Kings cut back seriously the budget.

            Then the manuscripts started to deteriorate, and in some decades the size of the Library went down to lesser dimensions, viable at the smaller budget. World’s Biggest Library could not exist too long as a relic. Without proper attendance & maintenance mice eat the books, say, in a generation.



            OK, in mid-3rd century AD (the time of Diogenes Laertius) there is no Great Library in Alexandria. Sure, there is a much smaller library, either with a Complete Aristotle, or not. A Complete Catalogue of Aristotle does exist.

            Diogenes Laertius' text is phrased as if the books still would exist. Maybe he does not know, where; but in the Empire smaller libraries operate in great number. However, surely, Aristotle's books are in scroll form. Remember: in 2nd century AD 99 % of Greek books are scrolls [6].

            In 3rd century the codex form starts even for Greek texts. As O'Donnell writes: "Books that made that transition [to codex] successfully had a reasonable chance of surviving and being read in the centuries to come, while books that did not were more likely to be orphaned. ... the plays of Menander, for example, which were not copied into codex form, were almost entirely lost to us..."

            Surely there were budget problems for research institutes and high schools. The Peripatetics copied the works used for teaching; this is may have been the Corpus Aristotelicum. The works of less "practical" usage remained in scroll form. Scientists and scholars could read (for a while) the old scrolls. Then came the troubles: the Migration on the West, Arabs in Egypt, social problems on the East. The List of the Victors in the Pythian Games and The Constitutions of 158 States were anyways obsolete, with the advent of Christianity and Global Monarchy. But why the 8 books of Anatomy were not copied into a codex?

            And why all Plato have been copied?



            This work obviously could be continued, and maybe it would be worthwhile. My present goal was simply to call attention at some points. The topics clearly needs interdisciplinary work, however maybe different persons could clarify different points further. If anybody has time and energy to continue, to me the important points are, e.g.:

A)    The simulation in Chap. 3 can be improved easily.

B)     Also, approximate formulae exist for the density of primes; exact ones do not. Also, some numerologist can easily count the primes between, say, 40,000 & 500,000; I do not have the proper codes.

C)    Palaeographists can easily tell if any hint exists for ancient scrolls with cca. 6000 lines (or 3000 double lines). While my estimations on length/width are based on extant papyri, obviously such an enormous scroll cannot be expected in intact form. Still, maybe Gedankenrekonstruktion may exist from fragments.

D)    Maybe somebody has an idea why Strabo is so uninterested in Musaeon and its scientists.

E)     The Lyceum’s later history is obscure. (I mean not the building, but the scientific community.)

Maybe along these lines the story of Lost Knowledge could be clarified somewhat.



            I tried to discuss this topics with several colleagues of mine, but without success. Maybe all of them are Platonist.



 [1]       J. Barnes (ed.): The Complete Works of Aristotle. Princeton University Books, Princeton, 1995

 [2]       I. Bekker:  Aristotelis opera. Reimer, Berlin, 1831

 [3]       Diogenes Laertius: The Lives and Opinions of Eminent Philosophers, transl. by C. D. Yonge. H. G. Bohn, London, 1853

 [4]       V. Rose: Aristotelis Fragmenta. Third Edition. Teubner, Leipzig, 1886

 [5]       ***: http://www.greece.org/alexandria/library/library15.htm

 [6]       J. Hannam: http://www.bede.org.uk/Library2.htm

 [7]       J. J. O'Donnell: Avatars of the Word. Harvard, Cambridge, Mass. 1998

 [8]       L. Casson: Libraries in the Ancient World. Yale, New Haven, 2001

[9]        B. Lukács: http://www.rmki.kfki.hu/~lukacs/FOMENKO3.htm

[10]      L. M. de Rijk (ed.): Dialectica Petri Abaelardi. Van Gorcum, Assen, 1970

[11]      I. M. Bochan'ski (ed.): Petri Hispani Summululae Logicales. Marietti, Vatican City, 1947

[12]      J. Donat: Summa Philosophiae Christianae; Logica. Innsbruck, Oeniponte, 1914

[13]      M. Maróth: Aristoteléstôl Avicennáig. Akadémiai Kiadó, Budapest, 1983

[14]      Strabo: Geography. Loeb Classical Library, Harvard University Press, Cambridge






My HomePage, with some other studies, if you are curious.