[C]oncurrency – e.g. multitasking – is hard, we all know that.

In the following post I analyze the economics of concurrency, using the example of a layered conversation with two members, and many concurrent threads occurring in overlapping time intervals.

(If you would think it a fun exercise, write up a comment about another topic of choice in multitasking – besides conversations, that is — and I’ll merge it into a generalized theory.  I already have that theory in the back of my head one way or another, and  social proof by induction is nice (beware the pun.) )


Handling n+1 threads of conversation with another person concurrently requires:

  1. excellent working memory, to generate shared implicit context,
  2. excellent verbal intelligence, to generate shared explicit context for ambiguity mitigation,
  3. precision in phrasing,
  4. parsimony in phrasing,
  5. shared, similar, experiences,
  6. unshared, differing, experiences,
  7. similar time preference

Fulfilling these seven requirements, it is possible to handle any amount of conversations at the same time, where the amount must not conflict with:

a) your working memory limitations – most people can maintain five to nine different chunks of data at the same time quite well – to generate implicit shared context, or,
b) the verbosity of speech you can mentally afford to invest in, to generate explicit shared context, or;
c) the precision of speech you can mentally afford to invest in – from fluffy-emotive to precise-systemizing – or;
d) the use of the absolute minimum amount of words necessary to convey your point precisely;

and converges on having:

e) experienced, and grown up with, overlapping and similar, as well as differing past life histories, and;
f) overlapping future planning horizons, and;
g) similarity in future time orientation.

So you see, handling any amount of ongoing conversations with the same person is a matter of fulfilling those requirements, and not putting oneself under too many restrictions due to acting, and having acted, unconstructively.


Now, the above part was about one quite specific use case. Can you think up more?

Head tips to Bernard Spil for the idea and Curt Doolittle for review.


Leave a Reply