Mnemonics in Ancient India

[Steve Farmer]

I'd like to add a correction to Shailendra Raj Mehta's interesting
recent post: George Thompson clearly wasn't claiming that mnemonic
devices weren't used in Vedic traditions. All that he said was that he
knew of no elaborate mnemonics in India based on the construction of
complex visual associations, like those found in Western 'memory
palaces.' Thompson obviously is well acquainted with Vedic mnemonics,
which are based on modifications of sounds and rhythms in canonical
texts and not on associations (as in the West) between sounds and
visual images.

The fact that a long string of famous memorists has existed in India
(as Shailendra shows) doesn't, moreover, imply that they were using
visual mnemonics -- or any formal mnemonics at all. Neuropsychological
testing of memorists in the past two decades has shown that extensive
practice on its own, even in the absense of formal mnemonics, can do
the trick. A good review of the literature is found in Charles P.
Thompson, Thaddeus M. Cowan, and Jerome Frieman, _Memory Search by a
Memorist_ (Hillsdale, NJ, 1993). Thompson et al. studied one famous
Indian mnemonist, Rajan Srinivasan Mahadevan, at Kansas State
University over a period of three years. Among his other feats, Rajan
could rapidly repeat from memory the first 31,811 digits of pi. As
Thompson et al. showed in a long series of experiments, the speed with
which Rajan made his recitation ruled out the possibility that his
method involved visual associations ('memory palace' type mnemonics).
In chapt. 8 of their book, moreover, Thompson et al. looked at
thirteen famous memorists and found that while a few of them *did* use
visual mnemonics (one was Luria's famous memorist, Shereshevskii), the
majority did not. In most cases, years of practice in rote
memorization did the job. In most cases, phenomenal feats of
memorization were limited to particular modalities: e.g., mnemonists
who could memorize long chains of numbers quickly often couldn't
recall textual data encountered once any better than 'normal'
subjects. In other cases, the reverse was true -- it all depended on
what kind of data the subjects had spent most of their time memorizing
over the years. Although some genetic component might be operative,
long and hard practice seemed to be the critical factor -- no matter
whether or any mnemonic method was used.

The fact that even ordinary memories can be trained to perform
spectacular feats of memory *without* the use of visual association
was demonstrated in a famous series of experiments by KA Ericsson and
his colleagues in the 1980s. For one brief overview, see Science 208
(1980), 1181-1182. Abstracts of several more recent papers by Ericsson
and his colleagues can be seen by typing in "Ericsson KA" in Medline
at http://www.ncbi.nlm.nih.gov/entrez/query.fcgi. Under ordinary
conditions, most people can repeat at most a list of seven random
numbers read to them once. In their early papers, Ericsson et al.
showed that with extensive practice even ordinary subjects could be
trained to repeat back error free lists of up to *80* numbers heard
once. Again, extensive practice in memorization was key -- NOT the use
of any elaborate mnemonics, whether of the visual 'memory palace' type
used in the West OR the kinds of oral transformations used by Vedic
reciters. Everything depends on neural plasticity and lots of
motivation and practice -- nothing more.

Ericsson's research potentially carries one major implication for
studies of Vedic recitation. Both native and European Vedicists (e.g.,
Staal), often claim that the reason for the extraordinary accuracy of
Vedic recitation has to do with the variety of complex mnemonic
devices used in training Vedic reciters. Staal, e.g., argues in _Oral
Tradition and the Origins of Science_ that manifold verbal
'modifications render the oral transmission even more firm and stable'
(referring to the kinds of modifications involved in
recitations of samhita, padapatha, kramapatha, jatapatha, ghanapatha,
etc.). One possible implication of Ericsson's work -- and I only raise
this as a possibility -- is that *no* mnemonics at all may be
necessary for 'tape-recorder-quality transmission': Extensive practice
alone may get the job done.  In China, so far as my Sinologists tell
me, phenomenal feats of memorization were achieved using only what
E. Bruce Brooks (of _The Original Analects_ fame) dubs tongue-in-cheek
as 'knuckle-cracking mnemonics.' Bruce writes (personal communication):

  The memory device is rote repetition, all
  day every day for hours and the usual thump of the knuckle as corrective
  (same device, by the way, with the Japanese Palace gagaku musicians, they
  say the pieces from memory, using note names, hour after hour by their
  father's bed in the morning as the father is sleeping, except that when
  they say a wrong solfa syllable, the father's knuckle comes down on their
  head).

The great feats of Vedic recitation, on this view, may not be due to
the types of mnemonics used, but to the extraordinary motivation that
arose ritualistically from *having* to get the sounds right -- since
if they were wrong, the order of things in the universe (or at least
the order of things in the reciter's life) was endangered. That would
keep reciters practicing even harder than reciters in the West or in
China, for whom the meanings of the texts were, at heart, more
important than the sounds (as is supposedly the case for Vedic
reciters). I've often suspected that the great feats of memory
reported in the West had more to do with rote memorization than with
'memory palaces,' despite the huge number of books written on that
topic. I suspect -- and it is just a suspicion at this point -- that
the same might be true of famous feats of Vedic recitation. To steal a
phrase from Brooks: I suspect that formal memory aids in premodern
cultures were often 'not so much an aid to memorization as an
auxiliary of mystification.' As Ericsson's work suggests, long and
hard practice may be all that is needed.

Steve Farmer