THE ARAB CONQUEST 153necessarily incorporated borrowings from their predecessors in their
pristine and uncontaminated form. Where comparison can be made,
for example between extant sections of the Muqtabis of Ibn J:layyan
and the equivalent parts of the work of Ibn Idhari, the latter can be
seen not only to have been highly selective in what he took from his
model, whose text he considerably abbreviated, but also to be capable
of being deliberately misleading. Ibn Idhari was only too willing
to distort the information that he had received, for example in
the interests of concealing Muslim defeats, such as that suffered at
the hands of the Asturians in the battle of Lutos in 794. Likewise, the
limited and defective manuscript tradition of many of these works
can conceal real problems of source criticism. Where variant manu-
scripts do survive, it is striking to see how substantial the textual
differences can be between them. This is not just a question of scribal
errors, but of the wholesale recasting of historical episodes, probably
carried out for literary or ideological reasons, or possibly to incor-
porate aspects of another author's version of the same story.19
Another distinctive feature of Muslim historiography, which consti-
tuted one of its strengths but was equally liable to be abused, was its
reliance on tradition or hadith. Thus a story in an Arab historical
narrative, especially in the earlier sources, would normally be accom-
panied by the name of the informant from whom the author had
received it. In turn that informant's source might also be given, and
so on in a chain stretching back to the originator of the account, who
might have been an eye witness of the events described. The provi-
sion of this chain of informants enabled the reader to evaluate the
story for which they vouched. This was particularly true of stories
relating to the life and above all the oral teaching of the Prophet not
recorded in the Qu'ran. The authority of the tradition was directly
related to the status of the first informant. If he were one of
Mul].ammad's Companions, his earliest followers from his days in
Mecca, greater credence would be pl<\ced on the account than. if it
originated with a figure less closely associated with the Prophet. The
premium thus placed upon the religious status of the source resulted
in some cases in the fabrication of hadith. A spurious chain of inform-
ants, linking back to a supposed originator of high repute, could
validate a story or piece of doctrine that the contemporary author
wished to promote. Such items then picked up and re-transmitted by
another writer could rapidly acquire canonical standing.^20 Although
this practice began in the recording of stories about Mul].ammad, it