Introduction

2Although very little work so far uses the term “diachronic ESP”, there is a certain amount of work on the diachronic analysis of specialized text, and this may be thought of as being, at least part of, the content of diachronic ESP. Moreover, just like any text, understanding a specialized text from a previous period requires knowledge of the context in which it was produced. Hence this is an area where linguistics proper and cultural studies both have their importance, and thus where both linguists and specialists of cultural studies have a contribution to make.

3Of course, specialized text is probably as old as writing. The earliest examples of written text, in the 3rd millennium B.C. were specialized in nature, and Swales (1997) points out that these display characteristics that we associate with specialized text today. Morini-Garcia (2000) also sees texts from the 3rd and 2nd millennia B.C. as “ancêtres du discours spécialisé” (2000: 565). This paper, however, will deal with ESP rather than LSP, and hence with a much shorter time-scale. The work available is fairly piecemeal, but the largest section of the jig-saw puzzle which has been tackled is that of the diachronic analysis of scientific text, so that is where I shall start. I shall look first at the contribution of sociolinguistics to the field. I shall then look at the work of Michael Halliday in this area. His profound and insightful comments are among the most interesting to have been made in this field. My own work on the development of the scientific article follows on from that of Halliday. I shall then consider the particular area of the development of medical discourse, where the work of Salager-Meyer is particularly important. In addition, I shall mention some significant books of edited collections of articles. I shall also look at the question of multimodality, where non-verbal features, notably images, form an integral part of the discourse. Finally, I shall mention some contributions in specialized translation, and in areas other than science.

1. A sociolinguistic approach

4Those who work in the area of the diachronic analysis of specialized text usually take Bazerman (1988) as the seminal study in this field. Half of the chapters in this book had appeared previously as articles, and the book covers a wide range of topics, including, for example, experimental accounts in the Philosophical Transactions 1665-1800, the controversy surrounding Newton’s theory of optics, and experimental reports in spectroscopy in Physical Review, 1893-1980. Bazerman’s approach is that of a sociolinguist, and so he is interested in the social situation in which the text is produced, and what the text reveals about it. In the wake of Bazerman’s book, a number of others appeared with a similar orientation. Gross 1996, like Bazerman’s book, is based on previously published articles; ten of the twelve chapters had appeared in article form. He also deals with a range of topics, and his interest is in the rhetorical structure of scientific writing. Atkinson 1999 deals specifically with the Philosophical Transactions over the period 1675-1975, and again his interest is sociolinguistic. Valle 1999 also deals with the Philosophical Transactions from a sociolinguistic point of view, over the similar period of 1665-1965, but restricts her study to the life sciences. Gross et al. (2002) consider scientific articles in English, French and German from the 17th to the 20th century. They are interested in style, presentation and argument structure. Although they have more in the way of linguistic analysis, their quotes and examples from non-English sources are given in English translation only, showing that they are basically interested in what is said rather than how it is said.

2. A systemic functional approach

5One of the most significant contributions to this area is that of Michael Halliday within the framework, of which he was the founding father, of Systemic Functional Linguistics (Halliday 2014; Banks 2005). In Halliday 1988, he takes as his starting point the earliest known technical text written in English, Chaucer’s Treatise on the Astrolabe, which dates from 1391, and which he never, in fact, completed. This is a manual of instructions for the use of an astrolabe, and was ostensibly written for his son, said to be ten years’ old at the time. It is then not to be confused with something like a research article, being more like a modern teenager’s “how-it-works” type of book (Banks 1995, 1997). According to Halliday, Chaucer’s text is “a kind of technical, perhaps proto-scientific discourse” (1988: 165). He bases this judgement on Chaucer’s use of technical nouns, complex nominal groups, events under study expressed as relational processes, and activities expressed as material or mental processes. In Systemic Functional Linguistics a process is the function typically encoded in verbs. Material processes are physical actions and events; mental processes are events of a cognitive type, and relational processes are relationships of an attributive, identifying or possessive kind. Halliday then moves on several centuries to Newton’s Treatise on Opticks, which he takes as constituting the beginning of scientific English proper. This was not published until 1704, but was probably complete by 1687, and followed the controversial reception of his article “A new theory of light and colours” published in the Philosophical Transactions in 1672 (Fara 2002; Hakfoort 1988; Shapiro 2001). Where Chaucer’s text had been instructions for use, Newton’s is a discourse of experimentation: “in place of Chaucer’s instructions for use he has descriptions of action – not ‘you do this’ but ‘I did that’” (Halliday 1988: 166). Experimental activity is expressed through material processes, with mental processes being used for observation and reasoning. In the experimental sections Newton uses clause complexes (in other approaches these would usually be called sentences) with numerous subordinate clauses, but in the mathematical sections the clauses are usually simple but with long complex noun groups.

6One of the most significant points that Halliday makes relates to Newton’s use of the nominalization of processes. Nominalization is a way of “packaging a complex phenomenon into a single semiotic entity” (Halliday 1988: 167). Moreover, this use of nominalization is closely related to questions of thematic structure and progression. The process will frequently be expressed in its verbal form towards the end of a clause complex, where it constitutes part of the rheme, and would normally be new information. It will then appear in its nominalized form at the beginning of the following clause, thus functioning as theme and given information. This use of nominalization is therefore an important resource in the argument structure of the discourse, and this importance has increased as time has gone by. It is

an essential resource for constructing scientific discourse. We see it emerging in the language of this period, when the foundations of an effective register for codifying, transmitting and extending the ‘new learning’ are rapidly being laid down. (Halliday 1988: 169)

7With this device a clause can express relationships between processes, such as “process a causes process x to happen”. This tendency has gone on developing up to the present day, and Halliday suggests the following schematic representation of this phenomenon.

8This does not mean that we will inevitably find the final form in contemporary texts. In fact, any of the forms could occur. What Halliday is suggesting is that over time there will be a tendency to find the later forms (Banks forthcoming a). These points are taken up again in Halliday (1993), where he gives an analysis of the final section of Darwin’s On the Origin of Species by Means of Natural Selection, published in 1859. Once again he shows how the resource of nominalization enters into the thematic progression, with nominalized processes backgrounded as theme and given information. Moreover, as nouns, these processes are presented as being fixed and stable, something whose existence cannot be challenged: “the nominalization picks up on the preceding argument and presents it in this ‘objectified’ form as something to be taken for granted” (Halliday 1993: 98). As Halliday has said elsewhere, “you can argue with a clause but you can’t argue with a nominal group” (Halliday & Martin 1993: 39).

3. The development of the scientific research article

9Following on from the work of Halliday, my own recent research has concentrated on the scientific research article. This is notably the case in the second half of Banks (2008) and Banks (forthcoming b). In Banks (2008) I used a corpus of articles from the Philosophical Transactions covering the period 1700 to 1980. The corpus consists of two articles for the year 1700, one from the physical sciences, and one from the biological sciences; this was repeated at 20-year intervals up to 1980, thus giving a total of 30 articles. Analysis reveals an overall picture where the physical sciences are experimental from the beginning of the period and continue to be so until the second half of the nineteenth century when mathematical modelling begins to be a major interest alongside experiment. The biological sciences were originally observational and it is only in the middle of the nineteenth century that experiment begins to be used; from then on, both experiment and observation are significant in the biological sciences. The analysis shows that in the physical sciences the use of passives increased from about 25% of finite verbs to 30% by the middle of the nineteenth century and has remained at that level since then. In the biological sciences, the rate was initially about 15%, and rose to 30% in the course of the twentieth century. The use of nominalized processes seems to have been always fairly common. In the physical sciences, their frequency is about one per 30 words of text until 1920 when there is a rapid increase to about one per 14 words of text. In the biological field, the rate is less stable but settles at about one per 30 words in the mid-nineteenth century, increasing to about one per 23 words in the twentieth century. A significant development in the course of the twentieth century in the use of nominalized processes is their use as modifiers (i.e. with an adjectival function). This never occurs before 1900, but becomes fairly frequent thereafter. There are even cases where a nominal group made up of a nominalized process as modifier plus head itself functions as a complex modifier within a further nominal group, as in the following example from a 1960 article in the Philosophical Transactions, where ionization front functions as a complex modifier in the nominal group A strong R-type ionization front structure:

A strong R-type ionization front structure with one shock is the one which is most likely to occur.

10Analysis of the thematic structure shows that themes relating to the experimental area (that is, the object of study, the experiment or observation and the equipment used) account for 75% of all themes in the biological sector throughout the period and in the physical sector until the end of the nineteenth century. From this point on the rate falls significantly, and is replaced by themes of a mathematical nature. Thus, before 1900, examples like the following from an 1860 article are the norm:

By the addition of a little water a small quantity of syncoretin was then thrown down, in order to carry down the last traces of the less soluble crystalline compound, in case any were still present.

11In the course of the 20th century, examples like the following, from a 1960 article, become fairly frequent accounting for over 30% over the themes:

Since z < 1 for all D-types, equation (2.11) shows that v1< c1< a1, so that all D-type ionization fronts move subsonically relative to the fluid ahead.

13The commonest type of theme by far in the Philosophical Transactions is the object of study, which rises from 32% to 57% over the period, whereas in the Le Journal des Sçavans the commonest type of theme (37%) relates to humans other than the author, most frequently the author of a book under review. This is consistent with a primary interest in an object of study in the English journal compared to a primary interest in humans in the French journal. An analysis of the process types of the finite verbs (Banks 2005, forthcoming c) shows that in both journals relational processes account for 30% of the finite verbs. In the Philosophical Transactions however, the most frequent process type is material, accounting for 35%, with verbal process (processes of communication) accounting for 14%. In Le Journal des Sçavans relational is the commonest process type followed by material process, 27%, and verbal process 24%. The incidence of relational process shows that description is important in both journals; however, physical actions and events, indicated by the use of material process is even more important in the Philosophical Transactions. The relative frequencies of verbal process show that communication is more important in the French journal than the English. Study of modality indicates that there is a significant use of deontic modality in Le Journal des Sçavans (18% of all modal expressions), whereas this type of modality is virtually absent in the Philosophical Transactions. This results from the presence of items from such fields as theology and law in the French periodical, while these fields are absent from the pages of the English journal. Overall the results of the analyses show that the linguistic features depend on the editorial decisions made by Oldenburg and de Sallo about the scope and content of their respective publications. Moreover these decisions are themselves comprehensible in terms of the historical context in which they were made. This is a striking illustration of the way in which a text is intimately related to the context in which it is produced, of which it then becomes a part, and without which it cannot be fully understood. For example, the historical situation in which Oldenburg finds himself leads him to produce a journal restricted to the scientific and technical fields, because that is what his potential audience were interested in, but this leads to increased use of material processes (usually expressing physical events, or recounting experiments), whereas de Sallo’s brief leads him to opt for book reviews covering the whole spectrum of disciplines of the time, and this leads to increased use of verbal processes (usually expressing what is said in the books under review).

14Shorter studies of the scientific article would include such contributions as Magnet (2000) which analyses the discussion sections of articles in the area of nutrition studies over a period from 1929 to the 1990s. She notes that the length of discussion sections falls from an average of 850 words before 1960 to 625 words during the 1960s; this trend is however reversed after 1990 when the length rises sharply to an average of 1300 words. There is an increasing use of quotes and the discussion section in the later part of the period “participe fortement à la stratégie de validation de l’expérience” (2000: 125).

4. Medical discourse

15Within the general scientific field, medicine holds a particular position. Salager-Meyer has studied a number of features of medical discourse from a diachronic point of view. These can all be considered to be more or less related to interpersonal aspects of the language. In Salager-Meyer et al. (1996), and Salager-Meyer and Defives (1998), the authors looked at hedging in medical articles from the nineteenth and twentieth centuries. They distinguish between shields (i.e. basically, expressions of epistemic modality), use of the passive, approximators, expressions of doubt, and emotionally charged intensifiers. They find that shields and expressions of doubt were rare in the nineteenth century but became common in the twentieth century. The opposite is true of approximators and emotionally charged intensifiers, which are common in the nineteenth but not in the twentieth century. The use of passives increases from the end of the nineteenth century onwards, with the period of most rapid change being 1930 to 1960.

16In Salager-Meyer (1999a, 1999b) she considers the question of citations in medical discourse over the period 1810 to 1995. In (1999a) she distinguishes between those references which are critical and those which are not. She finds that until 1930 critical and non-critical references occur with more or less the same frequency, but from 1930 onwards there is a rapid increase in the number of non-critical references from about four per paper to twenty per paper by the end of the period. In comparison there is only a slight rise in the number of critical references. She notes that “earlier papers adopted a much more critical tone than modern ones” (1999a: 19), but that before the twentieth century criticism was usually preceded by courtesy markers. In nineteenth century papers, criticism is scattered throughout the paper, whereas in the twentieth century, where it occurs, it is usually restricted to the introductory section. She comments on the importance of the sociological environment in determining these interpersonal aspects of the text:

[S]ocietal development is a determining factor in the changing of textual patterns, and that persistence and change in the social system are both reflected in the text and brought about by means of text, i.e. that determining factors of linguistic change are intimately linked to and brought about by the social, historical and cultural context in which discourse is produced. (Salager-Meyer 1999a, 26-27)

17The question of criticism is also raised in Salager-Meyer (1998), and Salager-Meyer and Zambrano (2001). They distinguish direct and indirect academic conflict in English and French medical discourse. These occur at the same rate in both languages during the period 1810 to 1930. In this period, criticism is provocative and highly polemical, though in the English papers the criticism is frequently preceded by courtesy markers, which is not the case in the French papers. From 1930 onwards criticism in the English papers becomes less personalized and more object-centred, but the French papers remain fairly aggressive until the late twentieth century.

5. Edited collections

18In recent years there have been a number of collections of papers dealing with diachronic aspects of specialized text. To continue the medical theme, Taavitsainen and Pahta (2011) contains twelve contributions all using Early Modern English Medical Texts, a computerized corpus of just over two million words, which is itself the second of three components which make up the Corpus of Early English Medical Writing. This second component covers the period 1500 to 1700 and includes general and specific treatises, medical recipe collections, health guides, and articles from the Philosophical Transactions. The contributions cover such questions as use of cognitive verbs, definitions, code-switching, controversy and advertising, the development of specialized discourse, and the expression of stance.

19Moskowich and Crespo (2012) is also based on a computerized corpus, the Corpus of English Texts on Astronomy; again this is part of a larger corpus: the Coruña Corpus of English Scientific Writing. As the title indicates it is devoted to texts in the field of astronomy and covers the period 1700 to 1900. The eleven contributions which make up the book cover such questions as use of adjectives, new nouns and specific vocabulary, complex predicates, the pattern N + PREP + V-ing, hedging and thematic structure.

20Diachronic aspects of specialized translation are considered in Duris (2008). This is not specifically about translation to or from English, so it is LSP rather than ESP. Only one of the nine contributions concerns English, and deals with early translations of Newton’s Opticks into French (Baillon 2008).

21My own edited volume (Banks 2010a) is, again, not totally devoted to English but four of the twelve contributions do deal specifically with English texts, covering topics such as referring to others in academic articles over the period 1700 to 1980 (Banks 2010b), continental drift in the nineteenth century (Château 2010), translations into French of Darwin’s The Descent of Man (Engel-Gautier 2010), and the use of illustrations in The Lancet, over the period 1823 to 1905 (Rowley-Jolivet 2010).

6. Multimodality

22The final item mentioned above falls into the domain of multimodality, a field of study that has grown rapidly over the last couple of decades, particularly in the wake of Kress and van Leeuwen (1996). Multimodality takes into account all aspects of communication, not just those that are verbal, and the area that has attracted most attention is the contribution made by visual elements in a text. In this article (Rowley-Jolivet 2010), the author shows that over the period 1823 to 1905, images in the Lancet, which were basically figurative, evolved only slowly and in erratic fashion. It was only towards the end of this period that visual conventions in the medical article began to be established. Magnet (2001) looks at the development of images in the area of nutrition studies over a seventy-year period up to the 1990s. She notes that the space allotted to visuals does not increase over this period, but that their nature has changed:

[T]he widespread use of computers has only made the production of data easier, quicker and has therefore led to the amplified treatment and display of quantitative measurements, while photography, a more expensive and less rewarding tool has almost totally disappeared. (2001: 72-73)

23Two of the contributions in my edited volume, Banks 2013a, are of interest in the field of ESP. Somerset 2013 analyses the illustrations in Arabella Buckley’s nineteenth century popularizations of the theory of evolution. My own contribution (Banks 2013b) considers the images in early issues of the Philosophical Transactions for the period 1665 to 1670, a subject which I had previously considered in Banks (2009). One of the interesting features revealed by this study is an apparent circular organization of the images on the page.

24While the literature on multimodality as such is vast, studies which treat the question from a diachronic standpoint are relatively rare. Similarly, there are many histories of the development of images in specialized documents, and while these might provide interesting background for studies in diachronic multimodality, they lack the linguistic component which would bring them into the field of diachronic ESP.

7. Translation and lexicology

25Translation has always been with us, and indeed the first translation of an academic article from French to English dates from 1665 and appeared in the first issue of the Philosophical Transactions, the original French having appeared in the first issue of Le Journal des Sçavans (Banks forthcoming d). The development of translation and translation techniques could be a fruitful area of enquiry for diachronic ESP. Duris (2008) and Engel-Gautier (2010) have been mentioned above. To these could be added Dury (2005) and Dury and Lervad (2008). These are diachronic studies of terminology in the context of specialized translation. In (Dury 2005), based on a corpus from the area of ecology, the author notes that terms do not conserve the same meaning either over time or from language to language. Thus a concept may evolve, but the term used to designate it will not necessarily be modified. Dury and Lervad (2008) is based on a corpus of texts in mineralogy (more specifically relating to petroleum), covering the period 1800 to 1960. They show how initial synonymic variation can lead to specialization.

8. Other than science

26All of the above may give the impression that diachronic ESP is a question of the diachronic analysis of scientific discourse. This is only because this is the area which has so far received the lion’s share of attention, but all other specialized areas are potential fields of study. Although excursions into these areas are rarer, the following might be cited as examples.

27Wagner (2003) considers the evolution of legal language in English. The legal language of England had passed from Latin to French by the thirteenth century, and gradually moved to English from the mid-fourteenth century when pleas in English were permitted. The result is that “current legal language is a mixture of Latin, Old English and Norman French” (2003:103).

28Taavitsainen (1994) compared religious texts from late Middle English with scientific texts from that date up to the late seventeenth century. She brings out the relevance of sub-genre differences. Underestimating sub-genre differences is something which Biber and Gray (2013) have also recently warned against.

29It is also possible to study the development of language in more restricted and exotic fields. Wozniak (2015), for example, looks at the evolution of the discourse of mountaineering from 1870 onwards.

9. Final remarks

30An overview of this type can never hope to be exhaustive, and I am sure that many readers will be able to think of work which might have been included, but which has not been mentioned here, partly, no doubt, a reflection of my own personal interests. With due apologies to any authors concerned, I hope nevertheless that this gives a fair picture of the current situation of diachronic ESP. As can be seen from the above, the area that has received most attention is the analysis of scientific text, but even here, what is available is sparse and scattered, and there is vast scope for further research. In areas other than science, there are no more than a few fairly isolated studies, so the field is wide open.

31I am one of those who believe that there is an intimate relationship between a text and its context. A text is produced in, and, in some senses, by its context, of which it then becomes a part. Thus the context produces the text, but the text then modifies the context, so that there is a continual cause and effect operating between the two. In studying diachronic texts this is felt all the more strongly as the contexts in which they were produced are no longer immediately available to us. This can be seen in the differences between early issues of the Philosophical Transactions and Le Journal des Sçavans mentioned above. In my work on late seventeenth century scientific writing, I have found it necessary to read up on the history, and particularly the history of science of the period. Thus this is an area which can combine both linguistics and cultural studies. Just as Petit’s seminal work on FASP (Petit 1999) opened up a research space where those working in ESP, but with an interest in literature could find a place, I would like to think that diachronic ESP could provide a research niche (Swales 1990, 2004) for those working in ESP but who have an interest in cultural studies.

32This approach also helps put many current features into perspective. For example, the present dominance of English as the language of academic science, and increasingly in many other areas including the humanities, seems so well entrenched as to be virtually a permanent fixture. So it is useful to remember that this is a relatively recent development, probably much more recent than most people realize. Ammon (2012), who studied the period 1880 to 2005, shows that while articles in scientific periodicals, which are written in English, account for over 90% of the total since the mid-1990s, they accounted for only 50% in 1960, showing that there has been a rapid increase in the use of English in the second half of the twentieth century. Going a little further back in time, in 1910, English accounted for only 30% of scientific articles published, and at that time German accounted for a higher percentage of the publications than English; English only overtook German in about 1930. So for a period of about twenty years in the early twentieth century, German, not English, could be seen as the dominant language of science (Gordin 2015). In the social sciences, the percentage of articles published in English increased from 48% in 1951 to 76% in 2005.

33Work in this area need not be restricted to a particular theoretical approach. As can probably be seen from the above, work that has been done so far has usually been within a more or less functional approach, and this type of approach seems well suited to the task; but this should not be seen as being to the exclusion of other possible approaches, for example cognitive. There are also data banks of texts such as Gallica, the website of the French Bibliothèque Nationale. This is a vast mine of texts, but usually in image form, so not immediately suitable for lexicometrical work. Where text analysis is involved this can be either manual or computerized. Computerized approaches require suitable corpora, and several are already available, such as those mentioned above. This type of approach can treat large quantities of data, but is restricted to features which the machine can recognize. A manual approach is more flexible, but can only deal with a limited quantity of data. I feel that computerized and manual analyses are suitable for rather different types of task and that both have a part to play.

34Hence this is a research field where some inroads have been made, but where there are still vast tracts of virgin territory. I would like to think that established researchers in linguistics with an interest in cultural studies, or vice-versa will find new avenues to be explored in this field, and that novice researchers will find a research space in which to exercise their talents.

I would like to thank three anonymous ASp reviewers, whose comments have enabled me to improve this article. It goes without saying that they are in no way responsible for any shortcomings which remain.