The influence of the Royal Society on 17th-century scientific writing

An outline of the language policy of the Royal Society leads to the hypotheses that it contributed to the development of a new linguistic profile and a homogeneous linguistic structure of scientific texts in the second half of the 17th century. These hypotheses are empirically tested on a corpus of ca. 76,000 words, which consists in equal parts of texts from the first and the second half of the 17th century. The research method follows the principles of multi-dimensional analysis. The hypothesis of the new linguistic profile is supported on all five dimensions, whereas the homogeneity hypothesis is only partly supported. The results of the study allow a tentative sketch of the long-term development of scientific writing. 1 The Royal Society and its language policy The intellectual movement which led to the foundation of the Royal Society (RS) started in the first half of the 17th century. On the authority of John Wallis, Thomas Birch (1756/1968, vol. I: 1) traces its beginnings to the year 1645 when “several worthy persons residing in London, who were inquisitive into natural, and the new and experimental philosophy, agreed to meet weekly on a certain day, to discourse upon such subjects”. The meeting on November 28th, 1660, when it was decided to follow the example of other countries and found “a college for the promoting of physico-mathematical experimental learning” (ibid.: 3), is usually taken as the foundation date of the RS. From the very beginning the RS took a vivid interest in the accessibility and the dissemination of research results: “It was resolved, that every member, who hath published or shall publish any work, give the society one copy.” (ibid.: 25). It was planned to set up a library, in which all publications handed in to the society should be kept. ICAME Journal No. 33 66 The RS also considered it its duty to contribute to the shaping of an adequate writing style for scientific publications. Texts submitted for publication had to have the approval of the president, and prospective authors could be asked to change whole passages before their works were accepted for publication. The first historiographer of the RS, bishop Thomas Sprat (1667/1959: 113), mentions two fields in which the RS contributed most to the advancement of learning, namely the experiments themselves and the language in which they were reported. The language policy of the RS is summarized in these words: a constant Resolution, to reject all the amplifications, digressions, and swellings of style: to return back to the primitive purity, and shortness, when men deliver’d so many things, almost in an equal number of words. They have exacted from all their members, a close, naked, natural way of speaking; positive expressions; clear senses; a native easiness: bringing all things as near the Mathematical plainness, as they can: and preferring the language of Artizans, Countrymen, and Merchants, before that, of Wits, or Scholars. In 1664, the RS set up a language committee, which was to meet once or twice a month and be chaired by Dr. Wilkins.1 Unfortunately, there is no record about the precise rules this committee set up “for improving the English language”. Their general tenor can be gleaned from Chapter V of the 1663 Statutes of the RS, which is entitled: “Of Experiments, and the Reports thereof” and reads as follows: “In all reports of Experiments to be brought into the Society, the matter of fact shall be barely stated, without any prefaces, apologies, or rhetorical flourishes.” Robert Boyle, one of the most eminent founding members of the RS, set out the principles of an appropriate writing style in his Proëmial essay. He advocates a “plain and unadorn’d way of Writing” (Hunter and Davis 1999–2000, Vol. 2: 33) with “expressions [which] should be rather clear and significant, than curiously adorn’d” (ibid.: 16), and he admits the use of foreign words only “when Custom has not only Denizon’d them, but brought them into request” (ibid.: 17). Against this background it can be expected that under the influence of the RS a characteristic discourse type developed with a largely homogeneous linguistic form. Since the approach to the study of natural phenomena promoted by the RS was expressly different from that of previous periods,2 it is also to be assumed that the linguistic structure of the texts produced by its members differed from that of earlier scientific writing. The influence of the Royal Society on 17th-century scientific writing 67 These two hypotheses about the homogeneity and the novelty of the linguistic structure of scientific texts produced under the auspices of the RS will be empirically tested in this study. 2 Previous research on scientific writing The language of Robert Boyle, the best-known representative of the RS, has been described in some detail in Gotti (1996). In the context of the genre ‘experimental essay’, Gotti (2001: 227–237, 2003: 227–241, 2005: 227–241) summarizes the characteristic features of Boyle’s language under the headings brevity, lack of assertiveness, perspicuity, simplicity of form, and objectivity. Gotti’s descriptions, which are based on Boyle’s New pneumatical experiments about respiration, contain plenty of illustrative examples, but no quantitative data. Members of the Scientific Thought Styles project in Helsinki studied various aspects of scientific writing in the Late Middle English and the Early Modern English (EModE) period. Taavitsainen and Pahta (1995: 525ff.) noted a “period style in the latter half of the seventeenth century” characterized by features expressing personal experience (first and second person pronouns, imperatives, modal expressions and past tense active verbs). Taavitsainen (1993) applied factor analysis to the texts of the Middle English sub-periods ME3 and ME4 of the Helsinki Corpus of English Texts (HC). Its input was a set of putatively emotional and interactive linguistic features (first and second person pronouns, interjections, direct wh-questions, private verbs). The focus of the study was on religious texts, and they proved a rather heterogeneous text category, whereas the two medical texts did not differ too much on the emotional scale, being placed at its lower end. The frequency figures of personal pronouns presented in Taavitsainen (1994: 332) allow the conclusion that in the EModE period medical as well as non-medical scientific texts developed a more personal interactive writing style, but formed a linguistically less homogeneous text category. Biber’s model of multi-dimensional analysis (MD analysis), although originally set up for the analysis of Present-Day English (PDE) texts, has been applied in studies of the diachronic development of scientific writing by Atkinson (1996, 1999), Biber and Finegan (1997), and González-Álvarez and PérezGuerra (1998). Biber and Finegan (1997) based their research on the science texts contained in A Representative Corpus of Historical English Registers (ARCHER). They come from the Philosophical Transactions of the Royal Society (PTRS). Starting with the period 1650–1699, the texts are grouped in 50-year periods, and each period is represented by ten texts of about 2,000 words each. In their 1997 study, ICAME Journal No. 33 68 where they investigated dimensions 1, 2, 3, and 5 (i.e. involved vs informational production, narrative vs non-narrative concerns, situation-dependent vs elaborated reference, non-impersonal vs impersonal style), they found that scientific writing became more informational, less narrative, more elaborated, and more impersonal after the 17th century. Whereas the first two developments proceeded unidirectionally, the other two were reversed in the 20th century. Atkinson’s corpus which he subjected to MD analysis also contains texts from the PTRS arranged in 50-year intervals, but his sampling technique was different, and the intervals do not coincide with those of ARCHER (for a comparative description of the sampling methods for ARCHER and Atkinson’s corpus cf. Atkinson 1999: 68–71). The chronological range of the corpus covers the years 1675 to 1975. The articles of the relevant years of the PTRS were assigned to the research fields ‘physical sciences and mathematics’, ‘biological/life science’, and ‘others’.3 Each category is represented in Atkinson’s corpus with the same proportion of texts as in the corresponding PTRS volumes. As a consequence of this stratification, as Atkinson calls this part of his sampling strategy, the research field ‘physical sciences and mathematics’ represents 50 per cent of his corpus, the field ‘biological/life science’ constitutes 37 per cent, and the rest comes from ‘others’. His whole corpus comprises 243,304 words, so that each of his seven intervals contains nearly 35,000 words.4 On the whole, the results of Atkinson’s MD analysis correspond to those of Biber and Finegan (1997). The development of scientific writing after the 17th century proceeds unidirectionally on dimensions 1 and 2, in a zigzaggy way on dimensions 3, 4 (overt expression of persuasion), and 5. But the details of the more complex developments on dimensions 3 and 5 are not in line with Biber and Fingegan’s description. In Atkinson’s corpus the reversal of the development towards more elaborated reference sets in already after 1825, and a second change of direction begins after 1925. His texts become less abstract between 1675 and 1725, before the development towards greater abstractness sets in, which is reversed again after 1925. The development of scientific writing between the 15th and the 17th century is the object of an article by González-Álvarez and Pérez-Guerra (1998).They investigated the development of the genres ‘science’, ‘fiction’, ‘drama’, and ‘letters’. Three of their four science texts are taken from the HC; the fourth, Lanfrank’s Cirurgie, was provided by the Scientific Thought Styles project. It is now accessible as part of the Middle English Medical Texts corpus (Taavitsainen et al. 2005). The scientific sub-corpora contain 12,726 (15th century) and 11,149 (16th century) words. The MD analysis of this corpus yielded interesting results for the development of scientific writing on dimensions 1, 3, and 5 from the The influence of the Royal Society on 17th-century scientific writing 69 beginning of the 15th to the middle of the 16th century.5 On dimensions 1 and 3, the science texts of this corpus showed a clear development towards more involvement and less elaborated reference. On dimension 5, a comparison of the relevant figures justifies the claim of a moderate development towards a more impersonal style.6 In an additional step, the authors compared their genre dimension scores to Atkinson’s for the year 1675, and this comparison led them to claim that on all three dimensions there was a unidirectional development from the 15th to the 17th century. Taken together, these studies support the hypothesis that the linguistic structure of scientific writing in the second half of the 17th century was different from that of earlier and later texts of the same genre and that on dimensions 1 and 3 this period even constituted a turning-point in the development of the genre scientific writing. 3 Corpus and research method The corpus on which this study is based consists of six science texts, which form two sub-corpora of three texts each. The texts of the earlier sub-corpus date from the first half of the 17th century (non-RS texts), those of the later sub-corpus were written in the second half of the 17th century by eminent members of the RS (RS texts). These are the texts in chronological order: William Barlow: Magnetical aduertisements (1616) Gabriel Plattes: A discovery of subterraneall treasure (1639) John Pecquet: New anatomical experiments (1653) Robert Hooke: An attempt for the explication of the phænomena, observable in an experiment published by the Honourable Robert Boyle (1661) Henry Power: Experimental philosophy (1664) Robert Boyle: Continuation of new experiments physico-mechanical, touching the spring and weight of the air, and their effects (1669) Barlow’s text consists of 12 chapters, the first eight of which are included here. This amounts to 15,615 words. The text by Plattes is about several methods of discovering and treating different kinds of minerals. It comprises 10,831 words and was analysed as a whole. ICAME Journal No. 33 70 Pecquet’s text is the English translation of his Latin original Experimenta nova anatomica (published in Paris in 1651). Several treatises are combined in this volume, and the passage included in the corpus is part of Experiments physico-mathematical of vacuity. It comprises 8,825 words. As the title specifies, Hooke’s text is a response to a report published by Boyle. The whole text was analysed; it contains 10,392 words. Power’s text is divided into three books “Containing new experiments microscopical, mercurial, magnetical”. The text chosen for analysis comprises all mercurial experiments and consists of 15,606 words. Boyle’s text contains 50 experiments and comprises about 63,000 words. The passage analysed here comes from the beginning of the text and comprises 15,238 words. Electronic files were produced from all texts. The textual bases were the facsimile editions of the texts by Barlow and Power,7 microfilm versions of the original texts in all other cases. All texts had to be keyed in manually, because the letters of the early prints are not recognized by the standard software of scanners. Table 1 maps the structure and size of the corpus. Table 1: Structure and size of the corpus The present corpus differs from the corpora of earlier studies in two ways. With its 76,507 words it is bigger than those, and the number of texts is bigger than in the relevant sub-corpora investigated by the members of the Scientific Thought Style project and by González-Álvarez and Pérez-Guerra, but smaller than the ‘science’ register in ARCHER and Atkinson’s modified version of ARCHER. Therefore it cannot be excluded that idiosyncratic traits of individual authors are reflected more than is desirable in the results of the analysis.8 The research method adopted here follows the principles of MD analysis. The choice of linguistic features was partly governed by pragmatic considerations (the corpus is not tagged, and at least some of the features should be automatically recoverable), partly by an effort to include those features which were found relevant in earlier MD analyses and which marked different dimensions. sub-corpus non-RS texts RS texts author Barlow Plattes Pecquet Hooke Power Boyle number of words 15,615 10,831 8,825 10,392 15,606 15,238