Developing the French academic periodical in the early 18th century: a systemic functional linguistics perspective

3.1 Comparison of generic forms in the two journals

The Journal de Trévoux was originally intended to be a monthly publication, but during its first year only appeared once every two months. By 1711, however, it had settled down to its intended periodicity and came out every month: thus, there were twelve issues in the course of the year. One innovation which had been established by 1711 was the final item in each issue which was headed Nouvelles Littéraires (‘Literary News’). This was a list of new publications from various places, with information and comments, but these comments were considerably shorter than the other items in the issue. These are grouped according to the places they come from; the issue for January has sections from d’Alemagne, de Lipsic (‘from Leipzig, Germany’), d’Italie, de Forli (‘from Forli, Italie’), de Venise (‘from Venice’), d’Angleterre (‘from England’), d’Amsterdam (‘from Amsterdam’), and de Paris (‘from Paris’). The generic forms of the items in the Journal de Trévoux are shown in Table 1.

As has been pointed out, this periodical is devoted essentially to book notices, and these do indeed constitute 75 % of the items printed during the year 1711. 16 % of the items are not book reviews, and do not bear any of the outwards signs of a letter, such as a salutation. I have called these articles, but some of these may well be extracts from letters without that being specifically made clear in the publication. The compilations of bibliographical information, of which there is one per monthly issue, make up 7 % of the items. There are a small number of items (2 %) which are definitely letters, since they have a formal salutation, and there is a single speech.

The Journal des Sçavans was a weekly publication and was produced regularly each week with, in 1711, a break from 14 September to 16 November. There were thus 43 issues in the course of the year. The Journal des Sçavans also had a final section in most issues headed Nouvelles de Littérature (‘News of Literature’), and like its rival these were divided into subsections according to place of origin (perhaps the Jesuits copied this feature from the Journal des Sçavans). The subsections in the January 1711 issue were de Ligourne (‘from Livorno’), de Leipsic (‘from Leipzig’), de Giessen (‘from Giessen’), de Francfort (‘from Frankfurt’), de Hall (‘from Halle’), de Copenhague (‘from Copenhagen’), d’Eysenach (‘from Eisenach’), d’Amsterdam (‘from Amsterdam’), de Franequere (‘from Franeker’), and d’Oxford (‘from Oxford’). The generic forms found in the Journal des Sçavans are shown in Table 2. The issues for each month have been grouped together.

Book notices constitute 81 % of the items printed in the Journal des Sçavans in 1711. The compilations of bibliographical information appear in most of the weekly issues, but not systematically, until the month of August. However, there are none in the month of September or after the autumn break. In total, they make up 18 % of the items in the Journal des Sçavans. There is a very small number of articles accounting for only 1 % of the items.

The results for the two journals are compared in Table 3. It can be seen that while both journals are largely devoted to book notices, this is even more strongly in evidence in the case of the Journal des Sçavans than that of the Journal de Trévoux. It might, however, be thought that the figures for the compilations of bibliographical information could have biased the results; since the Journal de Trévoux is a monthly publication, there are 12 such items accounting for 7 %, whereas since the Journal des Sçavans is a weekly publication (even though such an item was not printed in every single issue) they are considerably more numerous, and make up 18 % of the items. Had the Journal des Sçavans been a monthly, this figure would presumably have been much lower. If, however, the figures for bibliographic compilations are excluded from the calculation, book notices account for 80 % of the items in the Journal de Trévoux, but virtually all of the items (99 %) in the Journal des Sçavans.

Table 4 compares the figures for 1701 (Banks In press) and 1711. Over this ten-year period, the percentage of book notices appearing in the Journal de Trévoux has increased considerably, from 54 % to 75 %. This is at the expense of letters, which have almost disappeared, and, to a lesser extent, articles, which have fallen from 22 % to 16 %. In the case of the Journal des Sçavans, the percentage of book notices is the same for the two dates. At the same time, articles and letters have all but disappeared. To the extent that the compilations of bibliographical information deal with recent book publications, one could say that in 1711, 99 % of the items in the Journal des Sçavans deal with new books, so that it has become a publication devoted almost exclusively to newly published books. On the same basis, one could say that 82 % of the items in the Journal de Trévoux deal with new books, so it too seems to be on the way to becoming a journal devoted principally, if not exclusively, to new publications.

3.2 Comparison of contents in the two journals

The subjects covered by the items in the Journal de Trévoux are shown in Table 5. To avoid possible bias, the bibliographical compilations have been excluded. It can be seen that the subject which attracts the greatest amount of attention by far is Religion, which accounts for 36 % of the items, and even this figure may be underrepresented since History includes some items which are questions related to church history, and Law may sometimes be canon law. The second most frequent item is History, but, at 15 % of the items, much less frequent than Religion. These are the only two subjects which account for more than 10 % of the items. The third most frequent is Medicine, accounting for 7 % of the items.

The subjects dealt with in the Journal des Sçavans (excluding the bibliographical compilations) are shown in Table 6. Religion is again the most common subject, but in this case accounting for only 19 % of the items. Moreover, it is followed fairly closely by Law, Medicine and History, all in the narrow range of 15–17 % (Law 17 %, Medicine 16 % and History 15 %).

The results for the two journals are compared in Table 7. Religion is the most commonly treated subject in both journals, but whereas this subject accounts for 36 % of the items in the Journal de Trévoux, the equivalent figure for the Journal des Sçavans is only 19 %, so this subject is much more important in the Jesuit publication. History accounts for 15 % of the items in both journals, but whereas this is the second commonest subject in the Journal de Trévoux, it is only fourth commonest in the Journal des Sçavans, coming just behind Law (17 %) and Medicine (16 %). In the Journal de Trévoux, Medicine accounts for only 7 % of the items, and Law 4 %. So, Law and Medicine are considered much more important subjects in the Journal des Sçavans than in its rival publication.

In Table 8, the distributions of the subjects treated in the two journals for 1701 and 1711 are compared. In 1701, Religion was the most commonly treated subject in both journals, accounting for roughly the same percentage (25 %/24 %) of items in both. By 1711, this percentage has risen to 36 % in the Journal de Trévoux, but fallen to 19 % in the Journal des Sçavans. The only other subject accounting for at least 10 % of the items in 1701 was History, accounting for 10 % of the items in the Journal de Trévoux, and 16 % in the Journal des Sçavans. By 1711, this percentage has risen to 15 % in the Journal de Trévoux, but remained fairly stable at 15 % in the Journal des Sçavans. In the latter, however, two other subjects have caught up with History, and almost with Religion; these are Law and Medicine. In 1701, Law accounted for 5 % of the items in both journals. By 1711, this has risen to 17 % in the Journal des Sçavans, while remaining stable at 4 % in the Journal de Trévoux. Similarly, the percentage of items concerning Medicine, was virtually the same in 1701, 6 % in the Jesuit publication, and 5 % in the Journal des Sçavans. By 1711, this has risen to 16 % in the Journal des Sçavans, while remaining stable at 7 % in the Journal de Trévoux. Thus, the Journal de Trévoux has become even more focalized on religious questions, with over a third of the items published being about that subject. Of other subjects, only History can be said to be of any importance, but with less than half that accorded to Religion. The Journal des Sçavans on the other hand has reduced its interest in Religion, and by 1711 is producing a more balanced mix, with four subjects, Religion, Law, Medicine and History each accounting for between 15 % and 19 % of the subjects treated.

3.3 Selected sample

In order to look at some linguistic features in more detail, Banks (In press) took a sample of five articles from the Journal de Trévoux for 1701, all of which were categorized as being Religion, since this was the subject area most commonly dealt with. These texts yielded 674 finite clauses. The same has been done for the year 1711, taking the first five items labeled Religion, and this yielded 1,075 finite clauses. As in 1701, all were book notices. It will be noted that the items for 1711 are, on average, longer than those for 1701.

For 1701, in order to have a comparable number of finite clauses from the Journal des Sçavans, it was necessary to take the first 12 items dealing with Religion. This yielded 685 finite clauses. For 1711, the first 6 items dealing with Religion yielded 1,029 finite clauses. Again, all were book notices. It will again be noticed that the items selected for 1711 are longer than those for 1701, and this is much more the case for the Journal des Sçavans than the Journal de Trévoux.

The headings of the items selected from the Journal de Trévoux are given in Appendix A, those from the Journal des Sçavans in Appendix B.

4.1 The evolution of transitivity process types

Finite verbs are particularly important in that they constitute the node of the clause. I shall therefore look more closely at the process types encoded by the finite verbs in these texts. I identify five types of process: material, mental, relational, verbal, and existential (Banks 2019; Halliday and Matthiessen 2014). My interpretation of process types is semantic rather than grammatical (O’Donnell et al. 2008). Material processes are actions or events of a physical nature, involving concrete acts typically carried out by an Actor and affecting an Affected (or Goal), as illustrated in Example (1), where the process attaque (‘attack’) construes a physical action directed at a specific location.

Mental processes are cerebral events and may be cognitive, perceptive, or affective. Example (2) includes several cognitive mental processes in a series of verbs (a lû ‘read’, a examiné ‘examine’, a jugé ‘judge’) which construe acts of understanding and evaluation attributed to the Church.

Relational processes link two items, or one item with one of its characteristics. In Example (3), the relational configuration establishes the gallery as the object of discussion, while the clause qui étoit la premiere (‘which was the first’) assigns an attributive property to that location.

Verbal processes are processes of communication. Example (4) illustrates this category through the verbal process parlons (‘speak’), further extended by observe (‘notes’), both of which construe acts of scholarly discourse and reporting.

Existential processes are statements of existence. In Example (5), the existential marker il y en avoit (‘there were’) introduces the existence of a group of persons, whose quantity is then specified.

Those familiar with SFL will note that I do not use the category of behavioural process, which I find to be an incoherent category. The reasons for this are explained in Banks (2016).

The distributions of the various process types for the two journals are given in Table 9. Relational process is the commonest process type and accounts for about a third of the finite verbs in both (33 %/35 %). Verbal process is the fourth most common type with similar rates in both (18 %/15 %). Material process and mental process each account for about a quarter (24 %/23 %) in the Journal de Trévoux. In the Journal des Sçavans, however, the incidence of material processes is rather less (19 %), and that of mental processes rather more (30 %). This reverses the situation found in 1701, where the Journal des Sçavans had more material processes and less mental processes than its rival. Hence one can say that in this sample, the items in the Journal des Sçavans are more interested in intellectual matters and less interested in physical actions and events than its rival publication. It should be noted that the figures in Table 9 are those for the sample as a whole. This masks the fact that those for individual items may vary considerably. This is particularly the case for material process, which varies from 13 % to 35 % in the Journal de Trévoux and from 13 % to 30 % in the Journal des Sçavans.

The evolution of process types over the period 1701 to 1711 is shown in Table 10. It can be seen that for both publications the rate of relational processes has remained stable at around a third. In the case of the Journal de Trévoux, the rate of material processes has increased at the expense of verbal processes, with the rate of mental processes remaining stable. In the case of the Journal des Sçavans, it is the rate of mental processes which has increased, again at the expense of verbal processes, with the rate of material processes remaining stable. Thus, while over this ten-year period the development is not great, of the order of 6–10 % points, it does indicate a movement, in the case of the Jesuit journal towards a more factual stance, expressing physical actions and events, while the Journal des Sçavans has become more intellectual, with a higher rate of mental processes.

The Jesuit journal has thus maintained the process profile that it adopted at its creation, while the older publication, the Journal des Sçavans has evolved over this period, increasing its use of mental processes at the expense of material processes, thus concentrating more on intellectual matters. It is not impossible that this was in reaction to the creation of the Journal de Trévoux, but, while this is an intriguing possibility, the extent to which it was the case, if at all, must remain pure speculation.

4.2 Changes in thematic choice

The next linguistic feature I would like to look at is theme (or more strictly speaking topical theme). In SFL, theme is described as the speaker’s starting point, and in French, as in English, this is realized by the first major component of the clause (subject, circumstantial adjunct complement, or predicator) (Banks 2017b, 2019, 2025; Halliday and Matthiessen 2014). The clause is here construed as a main or coordinate clause with any associated subordinate or rankshifted clauses. This is strictly speaking a ranking clause, and for written discourse, some prefer the term T-unit (Fries 2001, 2002). I have classified the themes identified in the selection of articles according to the ad hoc categories used in Banks (In press). These categories are as follows. The author of the item in the journal itself or a group to which he belongs, including, where appropriate, the generic pronoun on (encoded Auth). In Example (6), where nous (‘we’) and the generic on (‘one’) both refer to the authorial voice of the journal item rather than to specific, identifiable individuals.

Example (7) designates author of a book under review, or in some cases the editor of older texts (encoded Auth/bk). In Example (7), the reference to M. de Meaux (‘Mr. De Meaux’) identifies the authorial agent.

Other humans or conscious creatures, including Jesus Christ and God (encoded Oth). Example (8) illustrates this category through references to Mr Faydit, who functions as third-party theme in the discourse.

Example (9) concerns Institutions (encoded Inst). In Example (9), l’Eglise (‘the Church’) is construed as an institution endowed with interpretive authority and moral responsibility.

Example (10) denotes the book which is under discussion, or part of it (encoded Bk). It illustrates this category through an explicit reference a chapter of the book which serves as the object of evaluation in the review.

Example (11) covers references to texts and documents other than the book under discussion (encoded Txt). In Example (11), ces manuscrits (‘these manuscripts’) refers to a set of documents and their condition.

Ment is used for references to mental processes, concepts, and argumentation. Example (12) exemplifies this category through a reference to the process of choosing (choisissant).

Example (13) concerns temporal references (encoded Temp). The prepositional phrase Au moment de (‘At the time of’) situates the proposition in time.

Example (14) designates Events (encoded Event). It presents a generalized event sequence framed as a causal or moral regularity.

Example (15) concerns Existential markers (encoded Exist). The pronoun il of the phrase il n’y a (‘there isn’t’) introduces the existence and quantity of an entity.

To these it has been necessary to add two extra categories to cater for a few items that could not be accommodated in the above. Example (16) concerns spatial locations or situations (encoded Sit).

Example (17) concerns physical objects (encoded Obj), a category occurring only in some Journal de Trévoux items.

In the selected items from the Journal de Trévoux, 495 themes were identified, and 370 in those from the Journal des Sçavans. The distributions of these categories for the two journals are shown in Table 11. It can be seen that in the Journal de Trévoux sample the three most commonly used themes are other humans, accounting for about a quarter (26 %), and the author of the book under review, and themes of a mental nature, both accounting for 18 %. In the Journal des Sçavans sample, we find the same three most common themes, but in this case the most frequent is themes of a mental nature, accounting for about a third (32 %), while other humans accounts for 21 % and the author of the book under review for 14 %. Hence the thematic highlighting (the author’s starting point for the clause) of mental themes is much greater in the Journal des Sçavans (32 %) than in the Journal de Trévoux (18 %). In contrast the Journal de Trévoux publication seems to have a greater thematic interest in people. Both other humans and the author of the work under review have a higher rate in the Jesuit journal. In the case of other humans, 26 %, compared with 21 % in the Journal des Sçavans, and in the case of the author of the book under review, 18 % compared with 14 % in the Journal des Sçavans. Moreover, if we take all categories relating to humans (Auth + Auth/bk + Oth), these total 52 % in the Journal de Trévoux, but only 42 % in the Journal des Sçavans. This gives a more intellectual feel to the Journal des Sçavans, with a particular interest in argumentation, while the Journal de Trévoux seems more down-to-earth with a greater interest in people.

In the following extract (Examples 18 and 19) from an item in the Journal de Trévoux, the thematic interest in humans is evident.

Each of the three themes in this short passage refer to historical characters who thus form the thematic centre of interest of the discourse at this point. The writer is interested in the history of Roman persecutions, and concentrates on the individuals involved. On the other hand, in the following extract, from the Journal des Sçavans, one can see that the themes are related to the construction of an argument.

It is now possible to consider the changes that have taken place in the choice of theme over the ten-year period 1701 to 1711. This is done in Table 12. In the case of the Journal de Trévoux, the two major changes seem to be the mental category and other humans. The use of the first of these has fallen from 35 % to 18 %, while the second has risen from 16 % to 26 %. In the Journal des Sçavans, use of the mental category has also fallen, and the use of other humans as theme has also risen, but the degree of change is much less than in the Journal de Trévoux. Use of the mental category has fallen from 38 % to 32 %, a fall of 6 percentage points, whereas the fall in the Journal de Trévoux is 17 percentage points. Similarly, the rise in the use of the other humans category is 3 percentage points, whereas in the Journal de Trévoux, it is 10 percentage points. Hence, over this ten-year period, the Journal de Trévoux seems to have moved away from a thematic interest in mental features towards an interest in people, while a basic interest in mental features has been maintained in the Journal des Sçavans.

There is a certain coherence in the fact that the Journal de Trévoux has moved to a position which favours material processes and the choice of humans as theme. Since theme frequently functions as subject, this means that they are frequently concerned with people doing things. In the case of the Journal des Sçavans, the move to a position using mental process verbs with mental features functioning as theme, reinforces the image of a periodical with an intellectual stance.

4.3 The alteration in the use of nominalization

Processes can not only be encoded as finite verbs; they can also occur as non-finite verb forms and as nominalized processes. Nominalized processes are a type of grammatical metaphor (Ravelli 1988; Taverniers 2003), and have been identified by Halliday (1988, 1998) as being particularly important in the construction of academic discourse. Nominalized processes can be categorized according to the process types they encode. Thus we may find nominalized material processes, as illustrated in Example (20).

Nominalized mental processes are exemplified in Example (21), where a cognitive activity is re-construed as a nominal entity rather than being realized through a finite mental process.

Nominalized relational processes are illustrated in Example (22). Here, the relational meaning of connection or association is realized through the noun la liaison (‘the link’), which nominalizes the relational process of “being linked” and allows the relation itself to be treated as an entity.

Nominalized verbal processes are illustrated in Example (23), where the act of saying or reporting is not encoded by a verbal process clause but is instead condensed into the noun la description (‘description’). This nominalization abstracts the act of verbalization and presents it as a textual product subject to evaluation.

Nominalized existential processes are possible, but none occur in the present sample. An example from the 1701 corpus used in Banks (In press) is given in Example (24). It illustrates how existence itself may be construed as a nominal entity through the noun l’existence.

The distributions of the process types of nominalized processes are given in Table 13. In the Journal de Trévoux, 318 cases of nominalization of processes were identified, and 260 in the Journal des Sçavans. This means that, on average, nominalized processes occur at a rate of one per 3.38 finite verbs in the Journal de Trévoux, and one per 3.96 finite verbs in the Journal des Sçavans, so the Jesuit publication uses this resource rather more readily than its older rival. The incidence of mental process and of relational process is virtually the same in the two publications. In the case of mental process this is a little over a third (36 %/37 %), and less than 10 % for relational process (6 %/7 %). On the other hand, the Journal de Trévoux uses nominalized material processes more frequently than the Journal des Sçavans, an incidence of 33 %, as opposed to 25 %, and verbal processes less frequently, an incidence of 26 % as opposed to 32 %. Hence, the use of nominalization in the Journal de Trévoux suggests that this journal is rather more interested in physical actions and events, while the Journal des Sçavans is more interested in what is said.

In considering the ways in which use of nominalization has altered over the ten-year period 1701–1711, it may first be noted that use of nominalization has reduced in both periodicals. In the Journal de Trévoux, the use of nominalization was one per 2.21 finite verbs, and this has reduced to one per 3.38 finite verbs in 1711. In the Journal des Sçavans, the rate was one per 1.93 finite verbs, and this has reduced to one per 3.96 finite verbs in 1711. So, use of nominalization is less in 1711 than it had been in 1701. The distributions of the process types of nominalized processes for the two dates are compared in Table 14. In the Jesuit periodical the distribution of the process types of nominalized processes has changed very little over this ten-year period. It is in the more established journal, the Journal des Sçavans, that changes have taken place. Nominalizations of actions and events are considerably lower, 25 % down from 35 %, while the rates of both nominalized mental processes and nominalized verbal processes have both increased, the former from 27 % to 37 %, and the latter from 26 % to 32 %. So in this periodical, what is done and what happens are nominalized less, while argument and what is said are nominalized more.

This is similar to the situation that was found for process types of finite verbs, with the Journal de Trévoux changing little, and the older journal changing much more, which is consistent with the conjecture that the changes were in response to the appearance of the new publication. Moreover, the position to which the Journal des Sçavans has moved shows a decrease in interest in actions and events, but increasing interest in mental and verbal processes. It may be remembered that in the early days of the development of SFL, verbal process was considered a subtype of mental process (Berry 1975).

4.4 A comparison of the distributions of finite verbs and nominalization

It may be of interest to compare the distributions of finite verbs and nominalized processes in these two journals. This is done in Table 15. Relational processes do not nominalize readily, so it is no surprise that the rate of nominalized relational processes is considerably less than that of finite verbs. Consequently all of the other process types have a higher rate of nominalized processes than of finite verbs. It may be noted however that the difference between the rate of finite verbs and nominalizations is slightly lower in the Journal des Sçavans for material and for mental processes, but considerably higher in the case of verbal processes. This confirms the penchant of the Journal des Sçavans for nominalizing verbal processes.