It was always thought that what the excavators found was the remains of Ashurbanipal's Library following its destruction in the fall of Nineveh in 612 BC. The tablets smashed as they fell to the floor. Once they are finally put back together we would have more or less complete tablets again, and they would constitute more or less the entire Library as it was. Researchers always wondered what lay hidden in the drawers of the British Museum. Would a rival scholar know of a fragment that would contradict their reconstruction of a text? Now the collection has been photographed and transliterated, the situation is very different. One significant conclusion that can be drawn is that these ideas are not correct, for several reasons.
There is evidence from archaeology indicating that the tablets were not found in the locations where they were kept in Ashurbanipal's day (see most recently George 2020). The two main findspots (Southwest Palace and North Palace) may not be as meaningful as we used to think. They do not necessarily represent two distinct "libraries" under Ashurbanipal. The specifics of their findspots within each location may also not tell us much about how they were kept by Ashurbanipal. Instead, they may reveal something of movements in the generation between his death and the fall of Nineveh, or processing or disturbance following the fall.
From the tablets themselves, a similar picture emerges. It is clear from several different text groups that we should not expect to reconstruct complete tablets. Only about 200 more or less complete tablets have been reconstructed over the last 170 years, and many of those are still missing large sections. Likewise, when you look at the extent of preservation of known texts, those tablets are typically preserved to only about 50%. Since the Library fragments have now been processed, we can no longer expect to find large sections of distinctive texts. The conclusion can only be that these fragments are not in the collection of the British Museum. They must still be at Nineveh. This indicates that the number of fragments currently known is not a reliable guide to the original size of the Library. It shows that the tablets must have been disturbed prior to excavation, otherwise the fragments would have been recovered. It also means that future excavations should discover many more fragments, and the context information generated by modern archaeological methods should help us understand much better what happened to the Library.
About 1,400 tablets or fragments contain a Library colophon. This is perhaps a surprisingly low number. Estimates of the size of the Library have often been around 5,000 tablets. This number should be viewed with caution. We do not yet have a clear idea of how many tablets did not have a colophon, although preliminary research suggests that the number of such tablets is low.
Joining of fragments could reduce the number of tablets represented by these 1,400 fragments. A trial with the most distinctive colophon type, however, suggests that the number of "joins" will actually be rather low. Just as many tablets are preserved only fragmentarily, so also the colophons are preserved only fragmentarily. Many fragments of colophons await discovery at Nineveh. We may have many tablets that once had a colophon, but that part of the tablet has not been excavated yet.
Estimates of the size of the Library have more recently been based on the assumption of an average of 5 or 10 fragments per tablet. The partial preservation of tablets forces us to reconsider such estimates. It must also be borne in mind that tablets are still being reconstructed. Recently, a large section of tablet has been reconstructed from numerous small fragments.
Dozens of fragments bearing Nabu temple library colophons have now been identified. They attest to a collection that should have been distinct from Ashurbanipal's own collection. It is unclear how many or which of the other fragments might also belong to this collection rather than Ashurbanipal's. It was clearly a much smaller collection, however.
Just as many different Ashurbanipal Library colophons have been found, so too there are different Nabu temple library colophons. These are the subject of a study by the project's MA student.
Around 300 tablets bear a colophon naming an individual scribe. This is a higher number than scholarship had expected. The names are a mix of Assyrians and Babylonians. What is remarkable about these tablets is that in almost all cases, each scribe is named only once. The exceptions are attested only a handful of times. Only one scribe stands apart from this pattern. He is Nabu-zuqup-kenu, a royal scribe from earlier times. More than a hundred tablets bear his name. Nabu-zuqup-kenu's tablets presumably were acquired by Ashurbanipal. His descendant worked for Ashurbanipal. Only a few tablets naming this man have been found, so the older tablets were probably not kept in his collection.
The fact that scribes are typically named only once shows that these tablets do not derive either from collections looted in antiquity by Ashurbanipal or from collections mixed in with Library tablets in modern times. It remains to be explained why so many scribes should have donated just one or a handful of tablets. They may perhaps have been part of the Nabu temple library, donated as a kind of ex-voto. Other explanations are plausible, however. The acquisition lists from 648 BC show a similar pattern, where many individuals listed there donated only one or a handful of tablets. These could simply be Library acquisitions. No individuals named in the lists can be identified in the colophons from Nineveh; admittedly, the lists are highly fragmentary. Some of the individuals named in the colophons can be dated to before Ashurbanipal's time; they can't be explained as coming from the 648 BC acquisitions. It is unclear when and how they entered the collection. The tablets with personal colophons are the subject of a dissertation by the project's PhD student.
Around 200 further tablets bear colophons that do not mention the Library, but do not record the scribe's name either. It is unclear to whom these will have belonged or how they came to the Library. Intriguingly, a plausible correlation can be made between one of these tablets and an entry in the acquisition lists; this is the first such correlation found.
A combination of textual analysis, palaeographic analysis, diplomatics analysis, and study of tablet groups, reveals a system behind the plethora of Library colophons. First, not all colophons are attested equally. There are about 300 examples of colophon Asb a, over 100 Asb b, and more than 400 Asb c/d. These together account for the bulk of the colophons. Alongside these are some colophons found less commonly, such as Asb l (more than 60), Asb q (more than 40), and smaller groups such as Asb g, Asb k group, and Asb r/s. Beyond these are about 20 colophons attested just once or in only a handful of examples.
The main colophon types show clear organisation within a system. They demonstrate that the Library was not the product of a limited burst of interest, nor a moment of intense looting. Asb a is evidently a property mark applied to tablets that belonged to Ashurbanipal prior to his ascent to the throne, and those written shortly thereafter. Asb b was the first Library colophon proper, written early in his reign; presumably in 660s BC. It is particularly revealing in the sources used by Ashurbanipal's scribes and the methods of tablet production at Nineveh. At some point subsequent to this, a more complex system was devised, by a more confident and ambitious king. Here, too, we can see chronological and intellectual development. There is a standard mature colophon, applied broadly across the Library. It displays some common variations, the significance of which has not yet been discerned. The extispicy texts were awarded a special colophon, Asb l. And several other text groups were then awarded their own specific colophons, derived from the original default colophon. For example, Asb q is reserved for medicine; Asb g for plant lists. These are the second group of colophons mentioned above. The third group must have specific explanations. An example is Asb m, which is found three times, all on manuscripts containing copies of the same archaic literary text. They presumably arrived together. It is not yet clear why these non-standard colophons were applied.
The tablets belonging to former kings would surely have become Ashurbanipal's property. Yet no tablets bear a colophon indicating that they belonged to any king other than Ashurbanipal. It is possible that some of the tablets with Asb a could have belonged in this category. This would be a small number. The texts detailing production under Esarhaddon suggest a much larger collection. Either they have not been found or they were copied and what was found were these copies.
The lists detailing acquisitions in 648 BC account for about 2,000 tablets. They would represent a large proportion of the ancient Library collection. But how large? It is plausible that more such lists once existed, accounting for yet more tablets. Among the surviving fragments from Nineveh, we have clear evidence for the existence of only about 2,000 tablets in total. The 648 BC tablets should therefore be prominent. Yet no-one has yet been able to show any surviving tablet to be one counted in the lists. Scribes named in colophons do not match those named in the list. And most surviving tablets clearly could not be among them, such as those with Library colophons. Our project has found one example where a tablet without named scribe in the colophon does seem to match one in a list. This suggests that we should look among such tablets for matches. This group is only 200 tablets strong though. Again, either these tablets have not yet been found or they were copied and what was found were these copies.
Given Ashurbanipal's strong and sustained interest in tablets, we are entitled to wonder how many other tablets might have been acquired. The later texts about an Assyrian king requesting tablets from Babylonia suggest that many tablets would have been involved. To what extent do these texts reflect reality? It is difficult to imagine that scribes did not send tablets in an attempt to gain favour, or that Ashurbanipal didn't make efforts to scour the southern cities for copies of texts that were not available in Assyria. It seems likely that at least some of the tablets with personal colophons belong in this category. We would expect many more. A plausible explanation would be that such tablets were copied, with the versions made to Library standards retained for the collection.
In the above paragraphs, copying has been suggested several times. Several points must be made. There are not nearly enough tablets to account for all the hypothesised copies. This suggests again that a significant part of the Library has not been found. A factor that requires consideration is the tablet copying methodology mentioned in Asb b. It states that multiple sources were collated in order to produce a single Library copy. Was this process retained later in the history of the Library? It is not clear what was done with the source tablets. They may not have been thought to belong to the Library proper. And few of them can be among the fragments found at Nineveh so far.
We noted above that we might look among the tablets with an individual colophon for matches with those in the acquisition lists. In particular we might look at tablets with such colophons that are also written in Babylonian script. It would make sense that such tablets might come from Babylonia. Some examples state that they are copies of tablets from Assyria; this might be interpreted in various ways. Other tablets in Babylonian script cannot be interpreted as deriving from the 648 BC acquisition. An extispicy text in Babylonian script bears a colophon naming Ashurbanipal's chief scribe, for example. It is not as simple as equating Babylonian script tablets with tablets acquired in 648 BC from Babylonia.
Fincke (2003-04) compiled figures describing the part of the Library written in Babylonian script. A similar exercise based instead on tablets with colophons reveals very similar figures. We might reasonably assume that these figures are therefore representative of the Library. Thus divination should comprise about half of the Library. About 20-30% should be magical texts. Then several other groups like lamentations, medicine, and word lists each comprise about 5-10% of the total. Some caution is needed though. There are significant differences between tablets with Library colophons and those with personal colophons. Divination is about twice as common in tablets with a Library colophon, while the other groups are half as common.
Jonathan Taylor
Jonathan Taylor, 'Making sense of the Library', Reading the Library of Ashurbanipal, Reading the Library of Ashurbanipal project, Department of the Middle East, The British Museum, Great Russell Street, London WC1B 3DG and Institut fuer Assyriologie und Hethitologie, Ludwig Maximilian University, Geschwister Scholl-Platz 1, 80539 Munich, 2022 [http://oracc.museum.upenn.edu/RLAsb/AssemblingtheLibrary/MakingsenseoftheLibrary/]