This may have been part of a general reduction in the use of graphs in mid-century Britain, apart from work in tides, but we do not address this issue here. For a recent history of the development of graphs in the nineteenth century, cf. Hankins, 1999.
In his Analysis of the Phenomena of the Human Mind (1829) James Mill equally made a lengthy argument to convince the reader that time could be represented by a straight line, an argument that was extensively criticised much later by the French philosopher Bergson (1889), who distinguished between time as perceived in our memory (historical time) and natural, or physical, time.
Similarly, it is perhaps one of Marcel Proust’s major tricks that at all stages of his À la recherche du temps perdu (note the title !) the age of Marcel is completely unspecified and impossible for the reader to figure out because the perspective change between the book’s protagonist Marcel and the narrator, Marcel Proust, is completely fluid.
And a successful one. Priestley’s chart went through several reprints. The accompanying introductory text, explaining the possibility of depicting the length of someone’s life by a horizontal line, disappeared after the first edition. The convention of a horizontal axis for time was not so well established even by the Jubilee volume of 1885 that Marshall could not argue, quite seriously, for a vertical time axis. We discuss this later.
Cf. Wood, 2000 for an overview and further references.
The first part of Stewart’s essay was first published 1815, the second, in French, in 1821. The full essay, together with John Playfair’s (the elder brother of William and professor of mathematics at Edinburgh) two essays on the progress of the mathematical and natural sciences, was first printed in 1824 as preliminary dissertation to the supplement of the 4th, and then 5th, and 6th editions of the Encyclopaedia Britannica.
Stewart,  1975, 26.
Stewart,  1994, 2, 235.
Ibid., 331-332. Stewart offered another intriguing reason to distrust statistical studies. In his eyes, these invariably offered the wrong policy advice. Where political economists rightly argued for the importance of free trade to improve the wealth of nations, statisticians « invariably encourag[ed] a predilection for restraints and checks, and all the other technical combinations of an antiquated and scholastic policy » (Works, 1994 3, 334). When statistical inquiry took off in Victorian England, this was, of course, no longer so. Indeed, Tooke and Newmarch, two of their more pronounced exponents highly welcomed the benefits of free-trade, which Newmarch compared in importance to the discovery of a new continent. Theodore Porter (1986, 5-6) emphasises how statistics as the « calculus of nature » was embraced by nineteenth century liberalism. The relation of statistics to political economy seems to have been easier on the continent. On political economists and statisticians in Germany in the early 19th century, cf. Nikolow, 1999 and 2001. On the reception of statistics in France, cf. Ménard, 1980 and Jovanovic, 2002.
Costigan-Eaves, MacDonald-Ross, 1990, quoting Playfair.
Costigan-Eaves, MacDonald-Ross, 1990, 324.
All quotes, Playfair, 1796, vi-vii.
To take money as token for wealth was not obvious. His contemporary August Crome, for example, took population as the relevant scale for his « maps » measuring the « strength of states » (see Nikolow, 2001).
On Mill’s early struggles with history, see especially De Marchi, 2002 (forthcoming).
Mill’s and Jevons’s views on the notion of disturbing causes are best outlined in Peart, 1995. A recent account of the distinction between contributing and disturbing causes in relation to the different explanatory strategies of econometrics and mathematical economics is given in Morgan, 2000, where she interprets these as modern versions, respectively, of Mill’s historical and deductive methods in economics.
In Logic (1843), Mill proposed a science of history based on his proposed science of ethology. Mill clearly played with the idea of doing justice to all contributing causes of historical events, thus explaining, for example « national character », where such a notion in the case of the more limited economic behaviour of individuals and nations was considered irrelevant. Mill never really embarked on the project, however, even though his empirical studies in economics show a clear awareness of the influence of what he considered disturbing causes for the abstract science of economics in concrete cases. For an exposition of Mill’s awareness, cf. Hollander, Peart, 1999.
Whewell,  1971, 282.
Whewell’s Philosophy of the Inductive Sciences first appeared 1840. We have used a reprint of the 1847 edition.
Whewell,  1967, 396.
For a history of tidology, cf. Cartwright, 1999. Whewell made pathbreaking contributions to the construction of cotidal maps. Many of Whewell’s phrases in tidology are still in use, like « age of the tide » and « luni-tidal interval ». Whewell read a paper on a mechanical graphical tide-recorder, constructed by T.G. Bunt to the Royal Society of London 1837. Cf. Whewell, 1838.
Hoover, Dowell 2001.
Even though debate has been recently opened up to what extent Mill subscribed to his own methodological views in his practical economic writing (Hollander, Peart, 1999).
In a commentary on a paper by William Guy on tabular analysis (Jevons, 1879, 657). We owe this reference to Jevons’s comments to Judy Klein.
Cf. Creedy, 1986 for an overview, cf. also Aldrich, 1987 ; White, 1989 ; Kim, 1995.
Cairnes, 1857, 86-88.
Jevons,  1958.
Stigler (1994) provides a definitive solution to the « puzzle » how Jevons estimated this equation. For an extensive analysis of Jevons’s use of the King Davenant Price Quantity Table in relation to the Theory of Political Economy, cf. Maas, 2001, 188-191.
Whewell,  1967, 397.
For a recent analysis of Jevons’s approach to index-numbers in his goldstudy, cf. Maas, 2001.
Jevons, 1884, 48, cf. also letter of 3 June 1863 to Cairnes, in Black, Könekamp, 3, 22-23.
The sunspot studies have been much analysed – cf. particularly Morgan, 1990 and Peart, 1996. Also Mirowski, 1984.
Ibid., 218. The date of publication of Milburn’s treatise, 1813, in the same period as Playfair’s work, may be noted.
A margin marking on Jevons (1884, 212), University of Amsterdam, Central Library, call mark 539 B 21. On Beaujon, cf. Stamhuis, 1989.
In ’The Solar-Commercial Cycle’, published in Nature, 26, 226-228, 6 July 1882. This short article is reprinted in Black, Könekamp, 1972-1981, 7, 108-112.
Black, Könekamp, 1972-1981, 7, 108.
His inferences were supported by both a re-estimation of the sunspot cycle and the finding of a small maximum in sunspots around 1797.
Black, Könekamp, 1972-1981, 7, 108.
Ibid., 1, 181, entry of 8. December 1861.
Ibid., 2, 450, letter to Richard Hutton, September 1, 1862.
Another aspect that might explain Newmarch’s dismissive attitude is the comprehensive character of Jevons’s undertaking. If Jevons showed Newmarch not only his plates, but also the general plan of the work, one can understand Newmarch’s hesitations. Jevons planned encompassing graphs for all four branches that formed the basis of the Statistical Society. Imagine your own reaction to a student entering your office with similar plans in view !
Marshall, 1885, 252.
We might interpret Clément Juglar’s « table-graphs » of 1889 in the same light – as a « natural » move to turn tables into graphs. Cf. Morgan, 1990, chapter 1, for an example.
The construction of causal arguments from graphs in the late 19th century and the switch from graphs to correlation and regression in economics around the turn of the century are both treated in much greater depth in Morgan 1997 who also gives a full account of Bowley’s work in these respects. Cf. also Morgan, 1990, chapter 2.
Cf. Klein, 1995 and 1997. Logical time, if we read her correctly, is unrelated to the « real » course of time. We can move back and forward on a curve in logical time which violates the « irreversible paths of change in temporal processes ». Hence, logical time is ultimately related to « static analysis » in economics, and not to the plotting of actual or imaginary data changing over time. Historical time, by contrast, involves irreversible patterns of change, whether these relate to actual data or to mental abstractions « drawn to follow a path of a variable over time » (Klein, 1995, 98-99).