Approaching the Journal des Sçavans, 1665–1695: a manual analysis of thematic structure

1. Introduction

A recent linguistics conference adapted Shakespeare by claiming that some texts are born digital, some texts achieve digitality and some texts have digitality thrust upon them. On the other hand, some texts are definitely not born digital, they certainly will not achieve digitality in the foreseeable future and they have a very high degree of resistance to having digitality thrust upon them. I shall call such texts “distinctly non-digital documents”. As an example of such a document, I would like to consider the Journal des Sçavans in the period 1665–1700. The Journal des Sçavans was the first academic periodical, and its first issue appeared, in Paris, on 5 January 1665, two months before the first issue of the Philosophical Transactions, in London. The Journal des Sçavans was founded by Denis de Sallo at the instigation of Colbert, who was Louis XIV’s first minister (Morgan 1928). Colbert perceived new knowledge, and new thought to be potentially dangerous to the centralized state machine that he was in charge of, and hence his objective in setting up the periodical was the state control of new knowledge. The cut-off point for my study is 1700, because, during this period, the Académie Royale des Sciences, which was founded in 1666 (Hahn 1971; Hirschfield 1981), published its findings only in luxurious editions which were often presented as gifts by Louis XIV to distinguished visitors (Licoppe 1994). It was not until 1699 that they took the decision to publish and distribute their work more widely, and the first issues of their publications began appearing in the early eighteenth century (McClellan III 2001). So from the beginning of the eighteenth century the situation changed radically, leaving the period 1665–1700 as one with peculiar, specific and interesting characteristics.

For this study, I shall use a corpus made up of five issues of the Journal des Sçavans for each of the years 1665, 1675, 1685 and 1695. Thus, the corpus has a total of 20 issues, with an estimated total number of words of 66,400. (I have a parallel corpus taken from the Philosophical Transactions with an estimated number of words of 77,300; thus, the total number of words for the whole of my corpus is 143,700, which, since the analyses are necessarily manual, is near the limits of feasibility. However, this paper will deal only with the Journal des Sçavans.)

The Journal des Sçavans is made up mainly of book reviews, which constitute 79% of the items in the corpus. Book publication was, at the time, the main method of disseminating new knowledge, so it seems a natural choice for those who wished to control that knowledge. Moreover, it covered the whole range of new knowledge. Indeed, if one wishes to control new knowledge, then it is sensible to cover all of it, not just a part. Thus, 17% of the items are in the domain of theology; this, in fact, underestimates the importance of religious matters in the corpus since historical items often concern church history, biography is sometimes hagiography and so on. History is the second most important domain accounting for 13% of the items, followed by medicine, accounting for 11%.

Early issues of the Journal des Sçavans constitute distinctly non-digital documents. It is true they are available on the Internet, but there all links with digitality come to an end, for they are available only in image form, which means that they cannot be tagged (unless someone is willing to undertake the onerous task of retyping the whole of the corpus into a computer). Similarly, they cannot be automatically read. The quality of seventeenth-century print is far too poor for automatic programs currently available (see Figure 1, which shows the first page of the issue for 23 February 1665). In addition, the letter “s” is represented by “long s” in initial and medial positions, and “long s” is not recognized by current programs. The digraphs “ct” and “st” are represented by a single letter block where the two letters are joined. Examples of this can be seen in Figure 2, which is the second page of the 23 February 1665 issue. No doubt these difficulties could be overcome, but this would involve a long and costly project of computer program development. In the meantime, and for the foreseeable future, the only way forward for the linguist is that of manual analysis.

Figure 1.

Journal des Sçavans, 23 February 1665, page 1.

Figure 2.

Journal des Sçavans, 23 February 1665, page 2.

The linguistic feature of thematic structure has convincingly been shown to be the driving force of academic writing, and to having an important role in the structure of argumentation (Halliday 1988, 1994, 1998; Banks 2008). Consequently, in what follows I shall present some results from a thematic structure analysis of this corpus of the Journal des Sçavans, 1665–1695. The definition of theme which I shall use gives topical theme as the first major component (i.e. subject, predicator, complement or adjunct). This may be preceded by a textual and/or an interpersonal theme (Halliday 2004a; Banks 2005). For convenience, in what follows “theme” means topical theme unless otherwise stated.

2. Grammatical function

When the themes of this corpus are considered in terms of their grammatical function, as one might expect, the vast majority of themes function as subject, while the second largest group is those that function as adjunct. This is shown in Table 1.

While one would expect subject to be the commonest function, followed by adjunct, it might have been expected that subjects should account for rather more, and adjuncts rather less than is the case here. However, percentages for individual years show that these figures are stable over time; this is shown in Table 2.

It can be seen that for individual years the percentage of themes functioning as subject remains in the range 72–74%, and that for adjuncts it remains in the range 21–23%. The following are examples of themes functioning as subject:

1. Ces Vies sont escrites d’vne maniere tres-agreable. Elles ont esté d’autant mieux recuës qu’il n’y auoit rien sur ce sujet dans nostre langue. L’Autheur dit dans la Preface de cét ouvrage, qu’il ne l’a entrepris, que pour seruir d’instruction à vn jeune seigneur.¹

(Journal des Sçavans 12 janvier 1665)

[These Lives are written in a very pleasant manner. They have been all the better received because there was nothing on this subject in our language. The author says in the Preface to this work that he only undertook it to provide education for a young lord.]²

In this segment, the three subjects, Ces Vies, Elles and L’Autheur, function as theme in their respective clauses.

Predicators functioning as theme are rare, accounting for no more than 1% of the themes. Even this figure may be somewhat overestimated since the impersonal verb falloir in initial position has been counted as a predicator functioning as theme, as in the following example:

2. Il faloit examiner si la multiplication de ces sieges episcopaux au lieu d’estre utile à la religion, n’y seroit point prejudiciable, & si les peuples estoient disposez à recevoir avec respect ces nouveaux Prelats.

(Journal des Sçavans 25 avril 1695)

[It is necessary to consider whether the multiplication of Episcopal sees, rather than being useful for the religion, was not harmful, and whether the people were disposed to receive the new Prelates with respect.]

However, in this particular case, even if one were to consider that Il faloit constitutes a grammatical metaphor of modality, and thus is an interpersonal theme, the first major component would then be the infinitive verb examiner, and so we would still have a predicator functioning as theme. Otherwise, predicators functioning as theme are usually imperatives:

3. Maintenant voyons s’il est possible que le globe de l’œil ou le Chrystallin change de conformation pour voir des objets differemment éloignez; & supposons par exemple qu’un œil puisse changer de forme autant qu’il est necessaire pour voir avec la mesme distinction un objet à un pied de distance, & un autre à six pieds.

(Journal des Sçavans 30 juillet 1685)

[Now let us see whether it is possible that the globe of the eye or the crystalline lens changes constitution to see objects at different distances, and let us suppose, for example, that an eye can change shape as often as is necessary to see an object one foot away, and another six feet away equally clearly.]

Here, the imperative verbs voyons and supposons function as theme.

The incidence of complements functioning as theme is virtually negligible. The rare cases which occur are pronominal complements, where the grammar of French requires that these be placed in pre-verbal position:

4. Aussi les traite-t-il de digressions Physiques pour la plus-part; parce qu’elles sont en effet des questions de Physique plustost que de mathematique.

(Journal des Sçavans 14 janvier 1675)

[Moreover, he treats them mainly as physical digressions, because they are, in effect, questions of physics rather than mathematics.]

The complement pronoun les in 4 is an example of this.

Adjuncts are the other major category, after subjects, accounting for 22% of themes:

5. Dans la 2^e, il traite amplement de la guerre de Candie, dont il tire les particularitez de l’Histoire de Nani.

(Journal des Sçavans 5 mars 1685)

[In the 2nd, he deals at length with the War of Candia, from which he gives some special points relating to the history of Nani.]

Here the adjunct Dans la 2e functions as theme. However, a large number of the adjunct themes are clausal in nature:

6. Ainsi ayant fait apporter en sa presence un petit enfant du Comte qui n’avoit encore que trios ans, il le menaça de le faire mourir à ses yeux & de devenir son ennemy, s’il ne se faisoit baptizer avec son fils; ce que le Comte executa avec encore tout le peuple qu’il gouvernoit.

(Journal des Sçavans 16 avril 1685)

[Thus, having had one of the small children of the Count who was not yet three years old brought into his presence, he threatened to have it killed before his eyes and to become his enemy, if he did not have himself baptized with his son. This the Count did together with all the people that he governed.]

I shall return to this point below.

There are some cases of cleft structures, where the clefted item functions as theme. Halliday (2004a) calls this predicated theme:

7. C’est à ces reflexions pleines de detours, de peu de bonne foy & de chicane, que M. de Tournai répond dans ce livre.

(Journal des Sçavans 4 juin 1685)

[It is to these reflections, full of twists and turns and little good faith, that M. De Tournai replies in this book.]

These account for 2% of the themes.

Similarly, there are a small number of extraposed structures, where the extraposition matrix functions as theme. Thompson (2004) calls this thematized comment.

8. Ce n’est plus une chose rare que d’en voir qui entendent le latin, & qui écrivent avec une pureté digne de l’ancienne Rome: mais Mademoiselle Le Fevre a peu de compagnes de son application à la langue Grecque.

(Journal des Sçavans 11 mars 1685)

[It is no longer rare to see someone who understands Latin, and who writes with a purity worthy of ancient Rome, but Mademoiselle le Fevre has few female companions as competent as her in the Greek tongue.]

These also account for 2% of the themes.

I shall now return to the question of adjunct themes, and more specifically to clausal adjunct themes. We have seen that 22% of themes function as adjunct; however, of these, more than half, 52%, are clausal, that is, 12% of all themes are clausal adjunct themes. This is shown in Table 3.

Here it can be seen that the 52% clausal adjunct themes are made up of 31% finite clauses and 22% non-finite clauses. The following is an example of a finite clause functioning as adjunct theme:

9. Avant que le Roy Aufrede eut fondé cette Université sur le fin du neuviéme siècle, l’an huit cens soixante-dix-neuf, les lettres y avoient fleury pendant long-temps.

(Journal des Sçavans 9 septembre 1675)

[Before King Alfred founded this university at the end of the ninth century, in eight hundred and seventy nine, literature had been flourishing there for a long time.]

The following is an example of a non-finite clause functioning as adjunct theme:

10. … mais les irruptions des Goths, des Huns & des Lombards, les ayant constraint de les abandonner, plusieurs familles se retirerent dans les Isles, où ils bastirent des maisons dont la ville de Venise & les autres Citez maritimes qui en sont voisines furent insensiblement formées.

(Journal des Sçavans 5 mars 1685)

[… but the incursions of the Goths, Huns and Lombards having forced them to abandon them, several families retreated to the islands, where they built houses, of which the town of Venice and other neighbouring maritime cities were imperceptively formed.]

The number of non-finite clauses here includes a small number of gerundive phrases, such as the following:

11. En se congelant à mesure qu’elle entroit dans la mer (ce qui forçoit la matiére qui survenoit de nouveau à couler par dessus ce qui estoit congelé) elle avoit fait encore retirer ses eaux plus loin en arriere que la portée d’un mousquet & environ deux fois autant d’espace en largeur.

(Journal des Sçavans 4 juin 1685)

[In hardening as it entered the sea (which forced the matter which continued to arrive to flow over that which was already hard) it had forced the water to draw back further than the range of a musket, and about twice the width.]

When the percentage distribution for individual years is considered, it can be seen that the overall percentage of clausal themes is relatively stable, but that the distribution of this between finite and non-finite clauses tends to vary. This is shown in Table 4.

The total percentage of clausal types remains within the range 50–55%. However, the percentage of finite clauses varies from 24% to 40%, and that of non-finite clauses from 15% to 29%.

Hence, it can be concluded that in this corpus the majority of themes function as subject, as would be expected. There is also a relatively large proportion of themes which function as adjunct, and of these more than half are clausal in form.

3. Textual and interpersonal themes

Textual themes occur in 29% of ranking clauses, but only 3% have an interpersonal theme. This is shown in Table 5.

Textual themes are usually fairly simple linking words or phrases as in the following:

12. D’ailleurs il parloit beaucoup, & aimoit à dire ce qu’il sçavoit.

(Journal des Sçavans 21 février 1695)

[Moreover, he spoke a lot, and liked saying what he knew.]

The following is an example of interpersonal theme:

13. Il est vray qu’on a laissé glisser une infinité d’erreurs dans l’art des Talismans, & qu’on en a dit des choses si peu croyables, que cet art ne passe plus que pour une superstition, & on ne regard à present ces medailles que comme autant d’enchantement & de sortileges.

(Journal des Sçavans 11 mars 1675)

[It is true that an infinite number of errors have been allowed to slip into the art of talismans, and so many scarcely credible things have been said that this art is no longer thought of other than as a superstition, and these medals are now considered as so much sorcery and spells.]

It will be noted that this example has the form of an extraposed clause, and though interpersonal themes are not common, those that do occur frequently have this form.

4. Semantic categories

The themes have been analysed in terms of their semantic categories. The categories used are those that were developed in Banks (2008). These categories were originally developed for analysis of the Philosophical Transactions, and while it is true that if one started from scratch to analyse the semantic categories of themes in the Journal des Sçavans, it is possible that the categories might have been slightly different, it is nevertheless the case that the categories used work reasonably well for the Journal des Sçavans, and it is useful to retain these for purposes of comparison (although this will not be done in this paper). The results of this analysis are shown in Table 6.

It will be seen that the commonest category is that of Oth. (humans other than the author), which accounts for 37% of the sample. This is followed by Obj. (the object of study or discussion), which accounts for 22%, and Inter (references to other texts), which accounts for 16%. Since the “other humans” category is relatively important, this has been looked at in greater detail. It is found that the following groups occur as other humans:

The author of a book under review (Writ.)

A person mentioned in a book under review (or other item) (Txt.sub)

The editor or publisher of a book under review (Ed.)

The translator of a book under review (Trans.)

References of a general nature (Gen.)

A person introduced by the reviewer of a book (Pers.)

The distribution of these groups is shown in Table 7.

It will be seen that the vast majority are either the authors of a book being reviewed (Writ.), which account for 41% of the sample, or persons mentioned in reviews (Txt.sub), which account for 36%. In the case of authors of books being reviewed, the word autheur occurs frequently:

14. L’Auteur de ce livre rejette le premier de ces sentimens, il doute de la solidité du second, & il approuve fort le troisiéme: mais il en ajoûte un quatriémé fondé sur l’authorité des Conciles & des Peres, qui donnent le nom de Canon à une certaine regle, discipline, ou genre de vie qui est conforme aux Canons & aux Maximes que les SS. Peres nous ont laissées dans leurs Ouvrages.

(Journal des Sçavans 11 mars 1675)

[The author of this book rejects the first of these opinions, he doubts the solidity of the second, and he strongly approves the third; but he adds a fourth based on the authority of the Councils and the Fathers, who give the name of Canon to a certain rule, discipline, or lifestyle which conforms to the Canons and Maxims which the Church Fathers have left us in their works.]

Otherwise, proper names commonly occur with this function:

15. M. Catherinot dit qu’Vfano Capitaine de l’artillerie au chasteau d’Anvers observe qu’en Portugal on garde encore par curiosité quatre anciens Canons dont l’un est nommé Indien parce qu’il fut fondu dans les Indes. Il ajoute à cela le nom des autres pieces de cette nature les plus extraordinaires que l’on a faites, comme celle qu’on appelloit la Diablesse de Boldue & le Triquetraque de Rome conservé dans le Chasteaux S Ange, lequel tire 5. coups de suite. Il vient après aux differents pîeces d’Artillerie & aux peuples à qui elles doivent leur nom & leur invention; ainsi il remarque que les moscovites inventerent le mousquet, les Arabes la carabine que l’on nommoit Arabine, & les Italiens de Pistoye les pistolets qu’on qu’on appella d’abord pistolles.

(Journal des Sçavans 30 juillet 1685)

[M. Catherinot says that Ufano, Capitain of Artillery at Angers Castle observes that in Portugal four old canons are still kept out of curiosity, one of which is called Indien because it was cast in India. He adds to this the names of other similar pieces among the most extraordinary that have been made, like that called the Daiblesse de Bolue and the Triquetaque de Rome conserved in Castel San Angelo, which can fire 5 successive shots. The he comes to various pieces of artillery and the peoples to whom they owe their name and invention. Thus he notes that the Muscovites invented the musket, the Arabs invented the carbine which is called the arabine, and the Italians of Pistoye the pistols which were originally called pistolles.]

Examples of the category of persons mentioned in reviews (Txt.sub) often occur when the reviewer is summarizing a work, particularly works of a historical nature, which is the case in the following:

16. Henri IV. qui commença à regner en 1399. fut qualifié d’Excellente Grace. Edouard IV. en 1461. se fit apeler Haut & puissant Prince. Henri VIII. eut d’abord la qualité d’Altesse, puis celle de Majesté.

(Journal des Sçavans 21 février 1695)

[Henry IV, who began his reign in 1399, was called “Excellent Grace”. Edward IV, in 1461, had himself titled “High and Powerful Prince”. Henry VIII first had the title of “Highness”, and then that of “Majesty”.]

In items other than book reviews, this category is simply a person mentioned in the text, such as this example in an item adapted from the Philosophical Transactions:

17. M. Coxes vient d’en faire de nouvelles en Angleterre sur ce sujet, & il écrit qu’ayant tiré beaucoup de sel de Fougere, & qu’en ayant fait dissoudre une partie à l’air humide apres l’avoir fait secher, le reste de la lessive estant filtré devint rouge comme du sang pur.

(Journal des Sçavans 11 mars 1675)

[Mr Coxes has just carried out some new ones on this subject in England, and he writes that having drawn a large amount of Fougere salt, and having dissolved part of it in a humid atmosphere, after drying it, the rest of the solution became as red as pure blood when it was filtered.]

Examples in the general category account for 8% of the sample. These occur often in introductory sections, like the following:

18. Quelques-uns l’ont attribué à Gerson Chancelier de Paris; mais ils n’avoient pas pris garde que l’Auteur du livre de l’Imitation se donne plus d’une fois la qualité de Moine, & que Gerson ne le fut jamais. Peu de personnes ont suivi le sentiment de ceux qui l’ont attribué à un Chartreux, ou à un Thomas Prieur de Vendeshem, parce qu’il n’y a pas eu des raisons assez fortes pour l’appuyer.

(Journal des Sçavans 1 juillet 1675)

[Some have attributed it to Gerson, Chancellor of Paris, but they have not taken into account the fact that the author of the Imitation says on several occasions that he is a monk, and this Gerson never was. Few have followed the opinion of those who have attributed it to someone from Chartreux, or to a Thomas, Prieur of Vendeshem, because there is no sufficiently strong evidence to support it.]

The impersonal pronoun, on, frequently turns up with this function:

19. On nous donne tous les jours de nouveaux systemes sur les fiévres.

(Journal des Sçavans 15 janvier 1685)

[Every day we are given new accounts of fevers.]

The roles of editor and publisher were not clearly distinguished at this time, and, indeed, were often the same person. This category (Ed.) accounts for 7% of the sample:

20. Quoy qu’il en soit celuy qui a pris soin de nous donner icy les œuvres de Lactance, ce Ciceron Chrestien comme on l’appelle, nous assure qu’il les a reveües sur les anciennes éditions & sur cinq differens Mss. Il ajoût à la fin le livre de Lactance de Mortibus Persecutorum que nous devons depuis ces dernieres années aux soins de M. Baluze, & il y insere les notes de cet infatigable auteur.

(Journal des Sçavans 4 juin 1685)

[In any case, the person who has taken the trouble here to give us the works of Lactantius, the Christian Cicero, as he is called, assures us that he revised them according to the ancient editions and five different ms. At the end of Lactantius’ book De Mortibus Persecutorum, he adds that we have been indebted for several years to the careful work of M. Baluze, and he inserts the notes of this tireless author.]

The impersonal pronoun, on, again frequently occurs with this function:

21. On a fait en mesme temps trois editions de ce liure.

(Journal des Sçavans 12 janvier 1665)

[Three editions of this book have been made at the same time.]

Persons introduced by the reviewer (Pers.) account for 6% of the sample:

22. Pachimere y a travaillé autrefois;

(Journal des Sçavans 6 mai 1675)

[Pachimere worked on it at one time.]

Finally, quite a number of works reviewed are translations of ancient texts, so the translator (Trans.) sometimes occurs as theme. These account for 2% of the sample:

23. Mademoiselle Le Fevre a reparé heureusement ces deux défauts: elle y a ajouté quelques fragmens qui avoient esté ômis dans toutes les autres editions;

(Journal des Sçavans 11 mars 1675)

[Fortunately, Mademoiselle Le Fevre has corrected these two faults: she has added two fragments which were missing from the previous editions.]

The object of study or discussion is more often of an abstract nature rather than a physical object, such as la question and le fait in the following:

24. LA question estoit de sçavoir, si le Pape accordant vne dispense de mariage au second degré de consanguinité, peut legitimer les enfans venus auparavant le mariage.

Le fait estoit, que Charles Barbier auoit eu deux enfans de Barbe Barbier sa niepce, auparauant de l’avoir espousée.

(Journal des Sçavans 12 janvier 1665)

[La question is to know whether by giving a dispensation for a marriage at the second degree of consanguinity can legitimize children born before the marriage.

The facts of the case are that Charles Barbier had two children by Barbe Barbier, his niece, before he married her.]

Where the object of study or discussion is a physical object, this is more likely to be in an item other than a book review. The following occurs in an item adapted from the Philosophical Transactions:

25. Ainsi le sel estoit comme tartareux & essential; estant seché par grand feu il diminua beaucoup en poids, & devint plus blanc, parce qu’il y avoit auparavant de l’huyle & de l’acide.

(Journal des Sçavans 11 mars 1675)

[Thus the salt was like deposited extract; having been dried by strong heat it lost a great deal of its weight, and became whiter because it had been oil and acid.]

References to other texts have been analysed by type of text. This is shown in Table 8.

It will be seen that books under review (BR) account for more than three-quarters (77%) of the themes referring to other texts (i.e. texts other than the Journal des Sçavans itself). This can be the whole of the book concerned or part of it:

26. Ce liure est diuisé en trois parties. La premiere contient les Regles des Peres de l’Orient, comme de saint Antoine, de saint Macaire, de saint Pacome &c. La seconde contient celles des peres de l’Occident, de saint Benoist, de saint Colomban &c. Enfin la troisesme comprend les Regles des autres Peres de l’Eglise pour les Religieuses, comme de saint Augustin, de saint Cæsarius &c.

(Journal des Sçavans 9 mars 1665)

[This book is divided into three parts. The first contains the rules of the Eastern Fathers, like St. Anthony, St. Makarios, St Pacomius, etc. The second contains those of the Western Fathers, like St Benedict, St Columban, etc. Finally the third contains the rules of the other Church Fathers for nuns, like those of St; Augustine, St. Caesarius, etc.]

Otherwise, other books (Oth.bks) account for a further 13%:

27. DE quatre cent ouvrages differens que la Republique des lettres doit aux Femmes Sçavantes, & que’un curieux a pris plaisir de ramasser soigneusement, il n’y en a pas de plus hardy que celuy-cy.

(Journal des Sçavans 11 mars 1675)

[Of the four hundred different works that the Republic of Letters owes to erudite women, and which an inquisitive mind has taken the pleasure to carefully assemble, there are none bolder than this one.]

Some of these are virtual, in that they are books which have not yet been written, books which it might be thought desirable that they should be written, and so on:

28. En attendant que cet Auteur traite dans un plus gros livre de toutes les autres parties de la Navigation, il en explique icy les principes …

(Journal des Sçavans 16 avril 1685)

[While we wait in the expectation for this author to deal with all the other parts of navigation in a larger book, he here explains the principle …]

The third category of references to other texts is that of texts other than books (Oth.txt), which account for 10% of the sample; these include letters, manuscripts and decrees:

29. … le manuscript est à la Bibioteque de Cluny …

(Journal des Sçavans 6 mai 1675)

[… the manuscript is in the Cluny library.]

Thus, books under review are the main type of other text thematized in the corpus.

5. Authors of book reviews

There is one semantic category of themes, which while it is not large, accounting for only 6% of themes, is nevertheless worthy of note. This is the category of authors (Auth.). Within this category, a group which is particularly significant is that of the authors of book reviews; these constitute 86% of all the authors in the author category. Part of the significance of this group is that the book reviews were written by the various editors of the Journal des Sçavans, or occasionally their close collaborators. It is interesting to see how these men referred (or did not refer) to themselves. The category includes references to the author directly, but also to groups of which he is a member. The results of this analysis are shown in Table 9.

The commonest way for the author of a book review to refer to himself is by the use of the impersonal pronoun on, with general reference (On.gen), where the co-text leads us to believe that the author is to be included:

30. On n’y trouve pas autant d’éloquence que dans les autres ouvrages de ce Père, parce que les sujets n’en sont pas capables: On en découvre pourtant assez en plusieurs endroits pour convaincre d’erreur ceux qui les ont attribuez à Eustathe Evesque de Sebaste qui est mort dans l’Arianisme, dont le caractere n’estoit nullement d’estre éloquent.

(Journal des Sçavans 6 mai 1675)

[One does not find here as much eloquence as in the other Fathers, because the subjects do not lend themselves to it. However, one does find sufficient in certain places to prove the error of those who have attributed them to Eustace, Bishop of Sebaste, who died an Arian, and whose character was to be anything but eloquent.]

The impersonal pronoun, on, is also sometimes used with specific reference to the reviewer himself. Examples with this function account for 10% of the sample:

31. … & on ne doute point que le monde ne fust bien aise d’apprendre dans trois feüilles de papier, ce qui demanderoit, sans ce secours, de années entieres. Mais on differera l’execution de ce dessein, iusqu’à ce que ce Iournal soit entierement estably, & qu’on ait trouué des personnes capables de bien traiter dans toutes les sciences, ces sortes de sujets.

(Journal des Sçavans 23 février 1665)

[… and we have no doubt that everyone was well pleased to learn from three sheets of paper what required, without this aid, years of work. But we will postpone the execution of this plan until this Journal is fully established, and we have found people capable of dealing with all fields and all sorts of subjects.]

The pronoun nous [we] accounts for 11% of the sample. Exactly half of these are inclusive nous, hence with general reference:

32. Nous avons encore un autre nouveau system sur les fiévres par M Minot Docteur en Med. Il se vend à Paris chez R. Pepie.

(Journal des Sçavans 15 janvier 1685)

[We have yet another account of fevers by M. Minot, M.D. It is on sale in Paris at R. Pepie’s.]

The other half of the nous examples refer specifically to the reviewer:

33. Nous avons parlé ailleurs de la premiere partie de cet ouvrage qui ne passe pas pour la meilleure piece de cet auteur.

(Journal des Sçavans 4 juin 1685)

[We have spoken elsewhere of the first part of this work, which can hardly be said to be this author’s best piece.]

On the other hand, the first-person singular pronoun, je, only occurs twice in the whole of the corpus, and the two examples occur in adjacent sentences:

34. J’AI parcouru dans l’onziémé Journal de cete année tous les ouvrages de feu M. Aubery. Je m’arrêterai maintenant un peu davantage à ce dernier, auquel il semble avoir raporté tous les autres.

(Journal des Sçavans 8 août 1695)

[I dealt briefly with all the works of the late M. Aubery in this year’s eleventh issue of the Journal. I shall now spend a little more time on this last one, in which he seems to have summarized all the others.]

In addition, there are a small number of occurrences of tout le monde [everyone], personne [no-one] and ce Journal [this Journal].

We can therefore conclude that the writer of a book review refers to himself rarely, and even when he does so it is likely to be in very general terms; in these cases, he is simply including himself in a very general reference, to the extent of being almost lost in the crowd. If we totalize the general uses of on, the general uses of nous and the occurrences of tout le monde and personne, we will have a figure for the overall uses of references by the writer of a book review to himself in general terms. By totalizing the specific uses of on, those of nous, together with je and ce Journal, we will have the corresponding figure for specific references by the reviewer to himself. This gives 81% for general references, and 19% for specific references. Thus, where the reviewer refers to himself, there is a four to one chance that this will be in general terms.

6. Concluding remarks

Our manual analysis has shown that in terms of the grammatical function of themes, 73% of themes function as subject and 22% as adjunct. However, of adjunct themes, 52% have clausal structure. Textual themes occur in 29% of ranking clauses, while only 3% have an interpersonal theme. When semantic categories of themes are considered, 37% are humans other than the author, and of these 41% are the author of a book under review. Objects of study or discussion account for 22% of the themes, and references to other texts account for 16%, of which 77% are a book under review. Only 6% of themes concern the author himself, and in the case of the author of a book review, even where he thematizes himself, in 81% of the cases it is only in general terms.

I trust that the reader will agree that these are interesting results, but beyond the inherent interest of the results themselves is the very fact that they have necessarily been produced by manual analysis. In these days of increasing use of computer analysis of digitalized texts, it seems useful to point out that interesting results can be produced by the manual analysis of distinctly non-digital documents. Since manual analysis will, for the foreseeable future, be the only means of studying the linguistic nature of distinctly non-digital documents, it is important that their study should not be eliminated simply because they are not amenable to computer analysis. The moral of the story seems to be that one should not throw out the manual baby with the digital bathwater.³

Disclosure statement

No potential conflict of interest was reported by the author.