68 Benjamin D. Sommer
is, tellingly, from Genesis 39. For the rabbis, any verse has signifi cant con-
nections to other verses throughout the canon, not just in the passage in
which the verse is found, and those cross-canonical connections produce
meaning. Of course, it is always possible that Genesis 22:1 is closely related
to Genesis 22:3, so that Genesis 22:3 holds the key to unlocking something
signifi cant in 22:1. But it is just as possible that a distant verse such as Mi-
cah 6:6 or 2 Samuel 7:18 or Psalm 110:4 or Proverbs 25:6 holds an impor-
tant key to Genesis 22:1; indeed, Genesis Rabbah cites all of these, but not
Genesis 22:3, to explicate Genesis 22:1. Th e midrashic interpreter will tend
to look to the verse further away, thereby demonstrating the Bible’s deep
underlying unity.
Th e fact that Genesis 22:1 appears in a Torah scroll next to Genesis 22:2
and not right next to the verses from Psalms, Deuteronomy, Isaiah, Ecclesi-
astes, Micah, Samuel, Psalms, and Proverbs that Genesis Rabbah cites when
explicating 22:1 is simply a result of the limitations of the technologies with
which humans write. It is not possible to put Genesis 22:1 next to all the
verses related to it when writing on a leather scroll or, for that matter, when
printing a book. Here we arrive at a crucial aspect of the midrashic concep-
tion of the Bible. For the rabbis, the Bible is not really a book at all. It is not a
scroll, and it is not a text (at least not in the way any other text known to hu-
manity in the fi rst millennium CE, and almost all the second millennium,
was a text). Rather, the Bible is a hypertext, a database with myriad internal
connections spanning the whole canon.14 Th ese connections link any one
verse to many other verses, which were in turn linked to a large number
of additional verses. Th us, a given verse had several literary contexts, each
of which implied several additional contexts. Th e physical data-storage
technologies available to humanity in the midrashic era (indeed, in most of
the postmidrashic era up until the late twentieth century as well) allowed a
verse to be physically contextualized next to only a few other verses. But in
reality (that is, reality as the classical rabbis conceived it), any one biblical
verse was part of a matrix of verses, each of which invoked additional ma-
trices.15 To be sure, one part (we might say, one row) of any verse’s matrix
was the context produced (or, rather, made obvious) by the written scroll
on which scripture was imperfectly recorded; this row links, say, Genesis
22:1 to 22:2 and 22:3 and might be referred to as the surface context or local
context. But that row, that surface context, was not necessarily the most im-
portant aspect of the verse’s matrix. Th inking of scripture as if it were text
in the way that Homer’s poems or Justinian’s laws are texts would severely