250 CHAPTER NINE
If, as seems to be the case, Weiss is suggesting that the pavement had a
single stable meaning, then he is necessarily supposing that its elements con-
stituted a kind of code, that the iconography had a fixed set of significations
readily accessible to all worshipers .This could be the case if we were to sup-
pose that the art was somehow closely connected to the ritual and liturgy of
the synagogues, so that these functioned to convey the meaning of the art to
the viewers .The supposition of a connection between art and liturgy is by no
means implausible and will be explored below .But it seems overwhelmingly
unlikely that the Palestinian liturgy of late antiquity was stable in any way .On
the contrary, the great period of synagogue construction was also, as far as we
can tell, a period of unprecedented liturgical dynamism, characterized in
some places by gradual rabbinization, probably almost everywhere by a ten-
dency toward professionalization, and the development and spread of thepiy-
yut, among other things; in any case, according to the consensual view, even
the rabbinic liturgy of the fifth and sixth centuries was characterized in Pales-
tine by a marked lack of fixity.^24 Not only did the liturgy vary from community
to community, but even within communities, prayer leaders were expected to
improvise their prayers .If the mosaics conveyed a single message to their
viewers, then that message must have been fixed by some means other than
the liturgy.
However, the variety of late antique synagogue decoration, the fact that
identical pavements have never yet been discovered, argues strongly against
the supposition that the art constituted a kind of iconographic code .Many
pavements have a roughly similar design, and many more use similar motifs
in distinctive and unpredictable ways .That the same are used individually to
decorate small objects and tombstones in itself implies the existence not of a
code but of a loosely constituted and unstable symbolic language, multivalent
or vaguely evocative rather than straightforwardly denotative .A useful coun-
terexample is the Mithraic tauroctony, the complicated scene of the god slay-
ing a bull found, with only minor variations, at the focal point of every
Mithraic shrine in the Roman Empire .The elements of this remarkably con-
sistent image obviously do constitute an iconographic code, known in princi-
ple to every initiate (if not to modern scholars), which almost certainly refers
to the central mysteries of the cult .Conversely, the stability of the Mithraic
iconography implies the essential unity and stability of Mithraic ritual.^25
(1981): 6–19 .For Goodenough (Jewish Symbols[1953], 1: 241–53), the Bet Alfa mosaic repre-
sents the ascent of the soul—a predictable reading but perhaps less implausible here than usually.
(^24) For general accounts, see S .C .Reif,Judaism and Hebrew Prayer(Cambridge: Cambridge
University Press, 1993), pp .146–52; I .Elbogen,Jewish Liturgy: A Comprehensive History(Phila-
delphia: Jewish Publication Society, 1993), pp .205–47; J .Heinemann,Prayer in the Period of the
Tannaim and Amoraim(Jerusalem: Magnes, 1984).
(^25) See Elsner,Art and the Roman Viewer, pp .211–21 .For the image, see M .Vermaseren,
Corpus Inscriptionum et Monumentorum Religionis Mithraicae, 2 vols .(Hague: Martinus Nijhoff,