Harvard University Press

Chapter Title: The Dark Ages of Probability

Book Title: Statistics on the Table

Book Subtitle: The History of Statistical Concepts and Methods
Book Author(s): Stephen M. Stigler
Published by: Harvard University Press. (1999)
Stable URL: https://www.jstor.org/stable/j.ctv1pdrpsj.16

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Harvard University Press is collaborating with JSTOR to digitize, preserve and extend access
to Statistics on the Table

This content downloaded from 37.166.22.103 on Mon, 08 May 2023 08:05:23 +00:00
All use subject to https://about.jstor.org/terms
 c ha pter 1 2

J
The Dark Ages
of Probability

The latter half of the seventeenth century might be termed the dark
ages of the theory of probability. The period began brilliantly enough,
with the works of Pascal, Fermat, and Huygens in the 1650s, and in
the end the darkness was permanently dispelled with the great treatises
by Montmort, Bernoulli, and De Moivre, all published in the decade
from 1708 to 1718. Between these bursts, all was apparently darkness
and silence. Yet, like the dark ages of history a thousand years before,
a closer look and new information can sometimes reveal faint glimmers
where none had been suspected.
What is widely regarded as the first printed work in probability the-
ory was a fifteen-page tract in Latin by Christian Huygens, appended to
a mathematics text published by Frans van Schooten in Leiden in 1657,
and reissued in Dutch two years later. Ironically, Huygens launched the
subject in a most uncertain way: its title was given as De Ratiociniis
in Aleae Ludo on the volume’s title page, and as De Ratiociniis in Ludo
Aleae on the tract’s first page. Huygens had visited Paris from July to De-
cember, 1655 (Brugmans, 1935, p. 40), and on that trip he had learned
of unpublished work by Pascal and Fermat on the problem of points (the
determination of a fair division of stakes in an interrupted dice game).
Huygens’s tract presented the solutions to fourteen problems in prob-
ability, solutions developed with some knowledge of the French work.
Pascal had written his own account in late 1654, including the success-
ful treatment of the problem of points, but it was not published until
1665, three years after Pascal died. This posthumously published book,
Traité du triangle arithmétique, gave a full discussion of binomial coeffi-
cients and their application to problems in probability. It is with Pascal’s
work that we first encounter the binomial distribution; he showed how
the binomial coefficients could be calculated easily in general and how
239

This content downloaded from 37.166.22.103 on Mon, 08 May 2023 08:05:23 +00:00
All use subject to https://about.jstor.org/terms
240 SOM E SE VE N TEEN TH- C EN TU R Y EX PLO R ER S

they could be used to evaluate probabilities. Edwards (1987) gives a full

account of Pascal’s work.

Probability in England
The published response to Pascal and Huygens was muted. Todhunter’s
(1865) encyclopedic history of probability mentions only three works
touching on probability between the work of Pascal and Huygens and
1690. Huygens’s tract was reprinted in 1670 by a Jesuit scholar named
Caramuel; a French mathematician, Sauveur, published some formulas
relating to the game of Bassette in 1679; and Jacob Bernoulli derived
two infinite series for a dice game in 1685 and 1690.
In the 1690s, however, there was a flurry of activity in England. Four
works appeared: a 1692 translation of Huygens by John Arbuthnot,
a 1693 note on lotteries by Francis Roberts (Roberts, 1694), and two
1699 works on the application of probability to testimony, one by John
Craig (Chapter 13), and one now known to be by George Hooper
(Hooper, 1699). It is natural to ask whether these publications were
part of an older English tradition or signaled an entirely new interest
in probability.
In fact, the English interest in chance and uncertainty has a long his-
tory. Official procedures at the Royal Mint requiring essentially random
sampling and a tolerance for variability go back to the twelfth century
(Chapter 21). In 1619, the Puritan divine Thomas Gataker wrote an in-
fluential treatise, Of the Nature and Use of Lots, attacking the notion
that the outcomes of lotteries and the throws of dice were direct signs
of God’s will, arguing instead that such chance events were determined
by natural, not divine, law, though their causes were not known to us.
Bellhouse (1988) gives a thorough discussion of Gataker’s work and the
reaction to it, placing it in the context of English Puritan casuistry in the
sixteenth and seventeenth centuries. Against this background, we might
expect that by the 1660s, English mathematicians would be receptive to
the works of Huygens and Pascal.

Richard Cumberland and the Laws of Nature

Van Schooten’s 1657 book was known in England soon after publica-
tion. For example, Isaac Newton purchased a copy in 1664 (Westfall,

This content downloaded from 37.166.22.103 on Mon, 08 May 2023 08:05:23 +00:00
All use subject to https://about.jstor.org/terms
 The Dark Ages of Probability 241

1980, p. 98), and Whiteside (1967–81, 1: 58–62) has reprinted three

pages of notes Newton apparently made at the time, showing that he
worked through Huygens’s tract carefully (and even extended it in mi-
nor ways). One indication that Huygens’s tract was noticed by other
readers can be found in a treatise on the laws of nature by Richard Cum-
berland, published in London in 1672. Cumberland, who was then in
Cambridge and later became bishop of Peterborough, wrote the book
primarily to refute what he took to be certain mistaken principles in
Thomas Hobbes’s philosophy. Cumberland was born in 1631 and died
in 1718. He was a remarkable early utilitarian philosopher who ar-
gued that the proper course was the pursuit of the greatest good, and
that random (“contingentis”) effects of human acts could be assigned
a value by following the rules for calculating chances and enumerat-
ing all the possible cases. Cumberland cited Huygens explicitly (1672,
p. 183; 1750, p. 283), and presented simple calculations involving one
and two dice as direct analogies to the evaluation of contingent effects in
agriculture, commerce, “and in almost all Employments, where human
Industry is concern’d and Busied” (Cumberland, 1672, pp. 322–323;
1750, p. 467).
Cumberland showed a natural tendency to reason inductively; he
argued that the observed tendency of animal species to live in a state of
peace rather than a state of hostility revealed a greater natural propensity
for peace than for hostility in animals (and, by inference, in man). The
argument was clinched by analogy: “For the Case holds here exactly the
same as in the Doctrine of Chances: it is more natural to suppose, that
Six will not turn up at the first cast of a Die, than that Six will; because
there are five possible Chances against such a Cast, and but one Chance
only for it” (Cumberland, 1672, pp. 119–120; 1750, pp. 193–194).
Again, Cumberland claimed that an individual’s pursuit of the com-
mon good (rather than the selfish behavior he saw Hobbes as promoting)
would more frequently benefit the individual, and hence was of greater
value: “Effects, caused by all other known Acts, are naturally contin-
gent, and, consequently, human Reason faithfully discharges its Office,
in case it directs us to choose the Event, which, in the general, must fre-
quently happen. For, a certain fix’d estimate of Value is put upon such a
Contingency as most frequently happens.”
Cumberland granted that in rare cases a man could benefit from the
use of force and treachery, but that did not render their use wise.

This content downloaded from 37.166.22.103 on Mon, 08 May 2023 08:05:23 +00:00
All use subject to https://about.jstor.org/terms
242 SOM E SE VE N TEEN TH- C EN TU R Y EX PLO R ER S

The Man, for Instance, who wagers, that, at the first Cast upon a
Pair of Dice, he will throw two Sixes, wins from him who lays an
even Wager, that, two Sixes do not come up first: Yet, notwithstand-
ing this lucky Cast, demonstratively true, (from the Nature, i.e. from
the Cubic Figure of a Die,) it is, that the Odds, advantage against
such a first Cast, are 35 to 1. And, consequently, the Chance of him,
who takes up the Wager, is to the Chance of him who lays the Wa-
ger, as 35 to 1, which are call’d the Odds. This Difference between
the Values of the Chances can and may, with great Propriety, be es-
timated and rated as the Chance, the Gain, i.e. The natural Reward
of the wiser, of a more prudent Choice. And, we are to determine, in
like Manner, concerning Damage or Loss, that, it is the natural Pun-
ishment of a foolish, of an imprudent Choice. (Cumberland, 1672,
pp. 322–323; 1750, pp. 466–467)

Thus Cumberland argued that decisions in life should be based upon

expected utility, and his analogy of a game of dice showed he had
understood clearly the fundamentals of Huygens’s tract.

Thomas Strode

Thomas Strode has been an unseen bit player on the stage of the history
of mathematics. If he is mentioned at all, it is usually as the recipient
of a 1676 letter from John Collins that was introduced as evidence
in the dispute between Newton and Leibniz on the priority for the
invention of the calculus (Whiteside, 1967–81, vol. 8). Yet it should not
surprise us that an age that could produce a Newton was a time of great
mathematical curiosity, and that in some cases this curiosity resided in
individuals capable of original work of their own.
Most of the little that is known about Thomas Strode’s life stems
from the fact that he attended Oxford University from 1642 to 1645
and thus came under the notice of Anthony à Wood, who in 1691–
1692 under the title Athenae Oxonienses published “An Exact History
of all the Writers and Bishops Who have had their Education in the
Most Antient and Famous University of Oxford.” Wood tells us that
Strode, the son of a Somerset gentleman, matriculated at University
College, Oxford, in 1642 at age sixteen, so Strode must have been born
about 1626. Strode studied under the Roman Catholic scholar Abraham

This content downloaded from 37.166.22.103 on Mon, 08 May 2023 08:05:23 +00:00
All use subject to https://about.jstor.org/terms
 The Dark Ages of Probability 243

Woodhead, and in 1645 Woodhead took Strode and another student

(Thomas Culpeper) on an extended tour of France. After four terms
abroad they returned to England, where Strode settled at Maperston
and “followed his natural Geny to Mathematics” (Wood, 1721). Strode
must have also studied law, for he was admitted to the Inner Temple as
a barrister in 1657 (Foster, 1892). Strode died sometime after 1690; one
source gives the year of his death as 1699.
Presumably Strode’s interest in mathematics was kindled by his tutor,
Woodhead, since Woodhead was widely known as an excellent instruc-
tor in mathematics. In addition to Strode and Culpeper, Woodhead was
specifically engaged as mathematical tutor to the duke of Buckingham
and to Lord Capell. Woodhead, however, appears to have published
nothing on mathematics, and among his students only Strode showed
a talent for original mathematical work (Culpeper did publish several
tracts on interest and usury).

Strode’s Treatise

Strode published only two works. One concerns probability: it appeared

in London in 1678 and is entitled A Short Treatise of the Combinations,
Elections, Permutations & Composition of Quantities; Illustrated by
Several Examples, with a New Speculation of the Differences of the
Powers of Numbers. The book is indeed a short treatise, comprising
only fifty-five pages, together with a two-page preface and an errata list.
Strode’s treatise was not unknown to his contemporaries; it is cited by
John Harris in a 1710 encyclopedia article on combinations. But the
treatise has been ignored by the history of mathematics, presumably
because as a work in probability it fell short of Pascal’s earlier treatise,
and as a work on combinations it was soon surpassed by a 1685 work
of John Wallis. It was brought to my attention by a dealer in rare
books. Todhunter was unaware of it, and of subsequent authors the only
reference to it of which I am aware is a passing notice by Edwards (1987,
p. 47). Yet Strode’s treatise deserves our attention for several reasons: as
an original work, as perhaps the first mathematical work on probability
in English, and as a picture of the level of understanding at the time.
Strode’s preface helps set the work in the proper historical setting; he
explained that the treatise was originally written in ignorance of Pascal’s
work, and it was only slightly revised to take account of Pascal:

This content downloaded from 37.166.22.103 on Mon, 08 May 2023 08:05:23 +00:00
All use subject to https://about.jstor.org/terms
244 SOM E SE VE N TEEN TH- C EN TU R Y EX PLO R ER S

Courteous Reader,
Above a year since, on the Entreaty of a very Worthy and
Publick-Spirited Friend, I gave my Consent that these Papers should
come to light; afterward understanding that some French Authors
had writ on this Subject, I put a stop to the Press; at length having
obtained those Authors, and perused what they have said concerning
the Argument herein handled, was willing (to prevent any abuses)
that the Press should proceed. I have since out of those Books
added two things; the first out of Malbranch alias the Author of
the Elemens des Mathematiques; namely, to give the number of the
several Compositions that may be made of 24 Letters, and that on
a double account, one to correct a Mistake of the Printer of that
Treatise; for I do suppose it to be no other, (for that the Number
of Figures, as also the first and the last are true;) the other, to
shew the manner of Operation which he hath omitted. The Second
is a Demonstration out of Monsieur Pascal’s Tract du Triangle
Arithmetique of what I had before by chance found out. Which you
will find in Page 33, but misplaced; for it should follow Page 42.1

The reference to Malebranche was in error, though the error was

a common one at the time. The book in question was in fact by Jean
Prestet (Robinet, 1967), and is not particularly noteworthy (it does
not discuss probability). The other statement, Strode’s claim that his
work was done before he saw Pascal’s Traité, is historically more
interesting. The claim seems plausible. The two works are done in
markedly different mathematical styles, and where they overlap, Pas-
cal usually goes further. Also, the portion of the argument on page 33
that Strode attributes to Pascal has the appearance of being a late
addition.
While Strode had not seen Pascal’s work when he wrote the greater
part of the treatise, he does appear to have known of Huygens’s tract.
He did not cite Huygens directly, but he did refer to “Francis Schooten
in his Miscellanies” in connection with combinations, which can only be
a reference to Schooten (1657). The other mathematical books referred

1. This preface and several other pages of Strode’s treatise are reproduced photographically
in Stigler (1988a).

This content downloaded from 37.166.22.103 on Mon, 08 May 2023 08:05:23 +00:00
All use subject to https://about.jstor.org/terms
 The Dark Ages of Probability 245

to are John Wallis’s Arithmetica Infinitorum (1656) and contemporary

work in progress by Thomas Baker.

Strode on Combinatorics

The first third of the treatise is concerned with counting. Strode dis-
cussed combinations, showing how the arithmetic triangle can be used
to determine their number. He referred to this as the Table of Figu-
rate Numbers. He showed how to compute the entries by multiplica-
tion without constructing the table. Of course, all of this and more
was in Pascal, but Strode was working at this point without the ben-
efit of reading Pascal, and we can at a minimum take this as testimony
that Strode was a capable mathematician. He gave a nonspecific refer-
ence to Tacquet, but he seems to have come upon the equivalent to the
formula
 
n n(n − 1)(n − 2) · · · (n − k + 1)
=
k 1×2×3×···×k
by himself. This is a significant achievement, although earlier mathe-
maticians had known of it (Edwards, 1987).
Strode next discussed elections. The number of elections of n quan-
tities is the number of nonempty subsets that may be formed; it is
given by
Xn  
n
= 2n − 1,
k
k=1

a result Strode attributed to Schooten. He then gave the rule for counting
“variations” (what we would now call permutations). He even treated
the case where not all of the objects were distinguishable. Strode did not
use abstract mathematical notation, but if we let Pkn be the number of
permutations of n distinguishable objects taken k at a time, then Strode
described how to compute
 
n
Pkk = k!, Pkn = × Pkk .
k
He even explained that, if the n objects included a subset of r that were
indistinguishable,

This content downloaded from 37.166.22.103 on Mon, 08 May 2023 08:05:23 +00:00
All use subject to https://about.jstor.org/terms
246 SOM E SE VE N TEEN TH- C EN TU R Y EX PLO R ER S
 
n
Pkn = Pkk /Prr ;
k
if there were R indistinguishable subsets of sizes r1, r2, . . . , rR , then
    Y
R
n
Pkn = Pkk Prrii .
k
i=1

In describing this last rule, he went a step beyond Pascal.

These discussions of methods of counting were accomplished
through and enlarged upon by a series of examples. Some were of a
routine nature: the number of “words” consisting of 3 letters from the
Latin alphabet of 24 letters is
 
24
P33 = 2024 × 6 = 12144;
3
the number of conjunctions of the (then) seven planets (presumably
including the earth and moon) is
           
7 7 7 7 7 7
+ + + + + = 27 − 7 − 1
2 3 4 5 6 7
= 120.
But some were a bit more involved. After asking what the value of
100 sheep offered in exchange for
 
100
barley corns
50
would be, Strode labored for three pages, using logarithms and the
latest data on the size of the earth to conclude “That if the Terrestrial
Globe should be covered with Guineys ten foot thick, there would not
be enough to pay for the Sheep: Neither if the Terrestrial Globe should
be converted to Barley, would there be Barley enough to satisfie for the
Sheep.”
Strode was imaginative in describing other large numbers as well. To
give the reader a sense of how large was the number of different possible
hands of 12 cards that could be extracted from a 52-card pack, he wrote,
“So that if 1000 men shall constantly deal 12 Cards for 12 hours each
day, excepting Sundays, they cannot deal all the several Games that are

This content downloaded from 37.166.22.103 on Mon, 08 May 2023 08:05:23 +00:00
All use subject to https://about.jstor.org/terms
 The Dark Ages of Probability 247

on 12 Cards in a Pack, in 54 Years, accounting that each man may deal

a Thousand Games in an hour.” The number Strode described,
1000 × 12 × 1000 × 54 × 6 × 52 = 2.02 × 1011
is not a bad approximation to
 
52
= 2.06 × 1011,
12
though it is smaller, as Strode stated.

Strode on the Chances at Dice

A portion of the treatise is concerned with dice, and some of the appli-
cations to the enumeration of chances Strode gave are more complicated
than can be found in earlier literature. The first mention of dice occurs
on page 15, where the rules I described above for permutations are illus-
trated. For example, for 3 dice the number of ways the outcome (1, 1, 2)
can be rearranged is 3; for 4 dice the outcomes (1, 1, 2, 2) and (1, 2, 3, 4)
have 6 and 24 possible “changes” or rearrangements; for 6 dice the out-
comes (1, 1, 1, 1, 2, 2), (1, 1, 2, 2, 3, 4), and (1, 1, 2, 3, 4, 5) have 15, 180,
and 360 different changes. After noting (p. 18) that the total number of
outcomes for a set of n dice is 6n, a result he gives for n = 2, 3, 4, 5 and 6,
Strode then proceeded to compute and table the probability distribution
of the total thrown with 2, 3 or 4 dice. He did this by enumeration, us-
ing his counting rules. He listed all possible totals, and for each possible
total he listed the ways the total could be obtained, up to permutations,
and summed those which corresponded to the given total to find the
number of chances out of 6n that give that total. Thus with 3 dice, there
are 3 ways to get a total of 6: (1, 1, 4), (1, 2, 3), and (2, 2, 2). There
are, respectively, 3, 6, and 1 distinguishable ways of obtaining these re-
sults, and so the chance of a total of 6 is 3 + 6 + 1 = 10 chances out of
63 = 216. The procedure was orderly, and it was carried out correctly.
Strode used his lists of possible outcomes to answer one further
question that was of interest to gamblers. Of the 1296 ways that 4 dice
could fall, he found that 216 gave double pairs of faces agreeing (or “In
and In” as he referred to the event). He found there were 360 chances
of no pairs, and thus 720 chances of a single pair.

This content downloaded from 37.166.22.103 on Mon, 08 May 2023 08:05:23 +00:00
All use subject to https://about.jstor.org/terms
248 SOM E SE VE N TEEN TH- C EN TU R Y EX PLO R ER S

In a two-page appendix (pp. 52–53), Strode discussed how the ap-

proach and results generalized to dice with other than 6 faces, and how
for some of the possible totals the chances could be found easily from his
Table of Figurate Numbers. In modern notation, Strode observed that if
a number n of s-sided dice were tossed, then of the total possible number
of outcomes s n, the number of chances that the total equals k is
 
k−1
n−1

as long as k < n + s. That is,

 
−n k−1
P (Total = k) = s for k = n, . . . , s + n − 1.
n−1

For s + n ≤ k ≤ (n − 1)s, he realized that more complicated enumeration

would be needed, but, for k > (n − 1)s,

P (Total = k) = P (Total = n(s + 1) − k).

Strode’s work on dice went beyond any that was published at the
time. Cardano, who died in 1576, left a short work that enumerated the
chances for 2 and 3 dice; it was published in 1663 (Todhunter, 1865,
p. 3). Galileo had likewise treated the case of 3 dice but, though Galileo
died in 1642, his work was only published in 1718 (Todhunter, 1865,
p. 5). Neither Cardano nor Galileo had generalized the rules for enumer-
ation as had Strode. Pascal and Fermat had gone well beyond Strode in
some respects (as in their treatment of the problem of points), but neither
had considered the general problem of the distribution of the total for
an arbitrary number of arbitrary dice, and their correspondence, while
dating from 1654, was not published until 1679. The tracts of Huygens
and Pascal showed a level of mathematical attainment capable of dealing
with Strode’s problem, but the interests of those authors lay elsewhere:
they wanted to know how to divide the stakes in interrupted games, and
how many throws of two dice would be expected to be required before
a 12 would occur. It could be argued that their problems were more dif-
ficult than Strode’s, but Strode’s problem of determining the distribution
of a sum has turned out to be historically more important.

This content downloaded from 37.166.22.103 on Mon, 08 May 2023 08:05:23 +00:00
All use subject to https://about.jstor.org/terms
 The Dark Ages of Probability 249

Strode on the Integration of Difference Equations

Strode presented one more “Mathematical Observation on Dice”

(pp. 22–24). Noting that, for 2 dice, the number of chances of obtaining
both dice below a number is the square of the number, he generalized
the result to more than 2 dice. In modern notation he had, for 2 dice,
P (both faces ≤ k) = k 2/62,
and, for m dice,
P (all faces ≤ k) = k m/6m.
This led him to an interesting question, and to what may be among
the earliest results on the integration of finite differences. For Strode ex-
plained in some detail a rule he had discovered inductively that enabled
him to use his Table of Figurate Numbers to solve problems of the type:
given any power and its differences, find another power and its differ-
ences. For example, given 33 = 27 and its (backward) differences,
∇33 = 33 − 23 = 19, ∇ 233 = ∇33 − ∇23 = 19 − 7 = 12,
and ∇ 333 = 6,
find 103. Strode’s solution was
     
7 8 9
103 = 33 + ∇33 + ∇ 233 + ∇ 333.
1 2 3
In general, we could write his rule as, for N > n,
X
k
(N − n)[j ] j k
Nk = ∇ n,
j!
j =0

where h[j ] = h(h + 1) . . . (h + j − 1) is the ascending factorial. It is

clear from Strode’s discussion that he came upon this relationship es-
sentially by integrating the difference equation ∇ k nk = k!. The formula
is of course a special case of the interpolation formula
∞
X
f (x + h) = (h[j ]/j !)∇ j f (x),
j =0

which is one of a class of formulas often referred to as Newton’s dif-

ference interpolation formulas. They are analogues for finite differences

This content downloaded from 37.166.22.103 on Mon, 08 May 2023 08:05:23 +00:00
All use subject to https://about.jstor.org/terms
250 SOM E SE VE N TEEN TH- C EN TU R Y EX PLO R ER S

of Taylor’s Theorem. Strode did not generalize his result beyond mono-
mials and their differences, but in that respect he anticipated later pub-
lished work by Newton. Other versions of this relation had been found
earlier, by Harriot, Gregory and Newton, but none of this work was
published by 1678. In published earlier work, Briggs and Mercator had
shown themselves familiar with predecessors of the formula; see White-
side (1961, pp. 232–252; 1967–81, 4: 14–73) and Goldstine (1977,
pp. 68–84) for those and other references. It would be interesting to dis-
cover if Newton was aware of Strode’s work.
Most of the remainder of Strode’s Treatise consisted of discussion
added after he had read Pascal. The promised addition starting on
page 33 presented a rule Strode credited to Pascal; we would write it
   
n n−k+1 n
= .
k k k−1

Strode next gave a Table of Figurate Numbers, a table of

 
n+m−1
m

for 0 ≤ n ≤ 30, 1 ≤ m ≤ 12; later in an appendix he added a table of

this for 31 ≤ n ≤ 100, 1 ≤ m ≤ 7. He then returned to a discussion of
Pascal. Strode reported that Pascal had given nineteen consequences
of the rule for constructing the table, and he repeated the five “choicest.”
In modern notation,
       
n n−1 n−2 k−1
= + +···+ ,
k k−1 k−1 k−1
       
n n−1 n−2 n−k−1
= + +···+ ,
k k k−1 0
    X n   n−1 
X 
n n n n−1
= , =2 ,
k n−k k k
k=0 k=0
   
n n−k+1 n
= .
k k k−1

Strode published one more book, a short treatise on “dialling,” ex-

plaining the use of certain geodetic instruments (Strode, 1688). That

This content downloaded from 37.166.22.103 on Mon, 08 May 2023 08:05:23 +00:00
All use subject to https://about.jstor.org/terms
 The Dark Ages of Probability 251

work was reprinted in 1697, and is referred to by Taylor (1954, pp. 224–
225, 405).

Conclusion
While Richard Cumberland’s treatise played a role in the development
of early utilitarianism, Thomas Strode seems to have had no impact
at all upon the development of mathematical probability. Those small
advances he made beyond Pascal were soon swept away by the full
emergence of probability as an area for mathematical research in the
eighteenth century (Hacking, 1975; Stigler, 1986a). If De Moivre knew
of the existence of Strode’s work, he did not think it sufficiently impor-
tant to require citation. Rather, the importance of Strode’s (and Cumber-
land’s) work for historians of probability is that they reveal that, even in
the “dark ages” of the end of the seventeenth century, ideas of apply-
ing mathematics to chance were circulating widely, and were accepted
without apparent resistance by mathematically educated people.

This content downloaded from 37.166.22.103 on Mon, 08 May 2023 08:05:23 +00:00
All use subject to https://about.jstor.org/terms