KINDS AND DATES 63

saga; and in other cases fragmentation will not account for the
wide differences between epos and ballad. The same is true of
Germany, where the ballads are even further removed from the
epics. Nor will fragmentation explain the historical ballads of
Spain, which are not fragments of long poems but songs essentially
episodic; nor explain the borrowing of adventurous matter from
abroad. What traditional epics can teach us is, chiefly, the rela-
tively late rise of the ballad genre. In Spain the 'cantares de gesta'
were still flourishing in 1344, when the scribes of the Second
General Chronicle copied in prose the text of a second Infantes de
Lara, unmistakably epical.1 The styles are successive, not con-
current,2 so that the mid-fourteenth century is indicated as a likely
moment for the rise of the Castilian ballad. Similarly in Scandi-
navia the Eddie poems come down to the eleventh century, and
they employ a rhymeless, alliterative technique quite distant from
the Vise' style. The rhymed, roughly syllabic, linguistically
medieval Viser' are thus clearly not older than the twelfth century,
at the soonest. The interposition of the early thirteenth-century
Thidrekssaga forbids placing a number of the epical 'viser' earlier
than then.

The epic evidence of Greece is more difficult to assess. We have
a clear allusion to ballad singers in the tenth century: 'accursed
Paphlagonians who put together songs about the experiences of
famous men, singing them at an obol apiece at each house.'3
Digenis Akritas has been identified as an Anatolian turmarch who
fell at Kopidnadon on the Euphrates frontier in 788. H. Gregoire
seeks to carry his epic back to the mid-tenth century, and repre-
sents the ballads as anterior and independent. The argument is
not entirely conclusive, but may serve to carry the ballads of Asia
Minor back at least to the twelfth century, if not sooner. We

1 See R. Menendez Pidal, Leyenda de los Infantes de Lara, Madrid, 1934.

2 W. C. Atkinson, in the Modern Language Review, xxxii, 1937, and Hispanic
Review, iv, 1936, argues that the epics flourished into the middle of the fifteenth
century, so that the epical ballads may be quite recently derived. One may say
that an epical ballad may be created by any minstrel who has a memory of an
epic text, that is to say, perhaps quite late; but there Is no such evidence of a
vigorous crop of new epics in the fifteenth century as in the Segunda Crdnica
General of the mid-fourteenth.

3 Arethas of Caesarea (850-932), cited by S, Kyriakides, and after him by
H. Gregoire. The latter's articles, 'Le tombeau et la date de Digenis Akritas*
and 'Autour de Digenis Akritas', Byzantion, vi, vii, 1931-2, are the principal
authority for Greek ballad and epic origins.