9
$\begingroup$

I was wondering where the term "hypergeometric" for the hypergeometric function $_2F_1(a, b; c; z)$ comes from. Wikipedia says that the term was coined by J. Wallis, but I couldn't find any (mathematical) reason why these functions are anything like hyper-geometric.

Does anyone know where this comes from?

$\endgroup$
4
  • 1
    $\begingroup$ mathworld.wolfram.com/HypergeometricFunction.html $\endgroup$ Apr 12, 2018 at 8:42
  • 2
    $\begingroup$ see also math.stackexchange.com/q/295258/442 $\endgroup$
    – GEdgar
    Apr 12, 2018 at 12:50
  • $\begingroup$ In Graham, Knuth, and Patashnik's Concrete Mathematics they have a comment when introducing the hypergeometric series on the etymology of the name in relation to the geometric series. I do not have the book in front of me right now. Maybe someone else can check. $\endgroup$
    – KCd
    Apr 12, 2018 at 13:21
  • 2
    $\begingroup$ On Page 206 (Second Edition) they write: "... is called hypergeometric series because it includes the geometric series as a very special case." $\endgroup$
    – p4sch
    Apr 12, 2018 at 16:14

2 Answers 2

8
$\begingroup$

This is the reason I believe: a geometric sequence is defined as a sequence where the ratio of two consecutive numbers is a constant $$ \frac{a_{n+1}}{a_n} = x\,. $$ You can generalize this notion by assuming the ratio to be any rational function of $n$ instead $$ \frac{a_{n+1}}{a_n} = \frac{P(n)}{Q(n)}\,. $$ Any such rational function can be factorized and rewritten as $$ \frac{P(n)}{Q(n)} = \frac{(n+a_1)\cdots(n+a_p)}{(n+b_1)\cdots(n+b_q)(n+1)}x\,. $$ If $Q(n)$ doesn't have an $n+1$ factor we can always say, for example, $a_p=1$, so we don't lose generality. With this parametrization it's easy to check that $$ \sum_{n=0}^\infty a_n = {}_pF_q(a_1,\ldots, a_p;b_1,\ldots, b_q;x)\,. $$ Hypergeometric then suggests that this is a generalization of the geometric sequence. More precisely $$ \frac{1}{1-x} = {}_1F_0(1;;x)\,. $$

Source: http://mathworld.wolfram.com/HypergeometricFunction.html

$\endgroup$
2
  • 1
    $\begingroup$ No, why? Setting ap=1 cancels the denominator $\endgroup$
    – MannyC
    Jan 4 at 7:26
  • 1
    $\begingroup$ OK (+1+1), I understand your intention now. Thanks $\endgroup$ Jan 6 at 13:45
2
$\begingroup$

The probability mass function of the hypergeometric distribution (related to the binomial distribution and the geometric distribution) is given by a ratio of products of binomial coefficients, just like the coefficients of the MacLaurin series of a hypergeometric function. This is my bet on the reason for picking the adjective hypergeometric for describing the $\phantom{}_p F_q$ functions.

$\endgroup$
3
  • $\begingroup$ It would be nice to know if the downvote is due to the fact that I am wrong, or something else. $\endgroup$ Apr 12, 2018 at 12:48
  • 1
    $\begingroup$ So the etymology is "geometric sequence" $\to$ "geometric distribution" $\to$ "hypergeometric distribution" $\to$ "hypergeometric function"? I found it slightly disorienting that the answer started in the middle of that chain rather than at one of the ends -- but not so disorienting that I'd downvote. $\endgroup$ Apr 12, 2018 at 12:59
  • 1
    $\begingroup$ @HenningMakholm: I am not sure that is the historical chain of events, but is sounds likely, doesn't it? $\endgroup$ Apr 12, 2018 at 13:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.