“God is invisible; that One created the world” — as we saw in the easy-read part 2, an apparently simple sentence like this actually folds two judgments (God is invisible / that One created the world) into one. What makes this compression possible is the relative pronoun. But in the sequel to Part II, Chapter IX, Arnauld and Lancelot show that a relative pronoun does not always work the same way. Ranging across Latin, Hebrew, Greek, and French, they tease out — example by example — the two roles hidden inside it: being a pronoun, and being a conjunction.
1. Two Faces, Three Combinations
A relative pronoun (qui, que, quem…) is actually doing two jobs at once, the authors say. One is standing in, as a “pronoun,” for a noun mentioned earlier. The other is joining two propositions into a single sentence, as a “conjunction.”
ce que nous avons dit des deux usages du relatif, l’un d’être pronom, et l’autre de marquer l’union d’une proposition avec une autre, sert à expliquer plusieurs choses dont les grammairiens sont bien empêchés de rendre raison.
(ToL working translation. Disclaimers on subsequent quotes are omitted; the same applies throughout.) What we have said about the two uses of the relative — one, being a pronoun; the other, marking the union of one proposition with another — helps explain a number of things grammarians have been unable to account for.
Usually both roles work together. But by patiently collecting real sentences, the authors find cases where only one role comes to the fore while the other nearly disappears. There turn out to be three types in all.
2. Type 1 — Keeping Both Pronoun and Conjunction Force
In the most standard type, a relative pronoun is both pronoun and conjunction. The authors take an example from Livy, describing Junius Brutus: he was “among (in quibus) the leading men of the city” who had heard that his own brother had been killed by their uncle.
il est visible que in quibus est là pour et in his ; de sorte que la phrase est claire et intelligible, si on la réduit ainsi : Quum primores civitatis, et in his fratrem suum interfectum audisset.
It is clear that in quibus stands there for “and among them” (et in his); the sentence becomes clear and intelligible if reduced this way: “when the leading men of the city [were there], and among them he had heard that his own brother had been killed.”
In quibus, in other words, carries at once the pronominal work of pointing to “the leading men of the city” and the conjunctive work of “and.” Without seeing this double role, the authors note, grammarians cannot properly parse the sentence.
3. Type 2 — Losing Pronoun Force, Keeping Only the Conjunction
But a relative pronoun can sometimes lose nearly all its pronominal force and become a bare conjunction, the authors say. They offer two examples.
The first is a turn of phrase peculiar to Hebrew. When the relative is not the subject of its clause but only part of the attribute — as in “dust that the wind blows away” (pulvis quem projicit ventus) — Hebrew leaves the relative only its second role (the conjunctive one), and expresses the pronominal role separately with a demonstrative pronoun, as if the relative weren’t even there: “the wind, it, blows it away” (quem projicit eum ventus), doubling the pronoun.
les Hébreux alors ne laissent au relatif que le dernier usage… et pour l’autre usage, qui est de tenir la place du nom, ils l’expriment par le pronom démonstratif… de sorte qu’ils disent : quem projicit eum ventus.
The Hebrews then leave the relative only its latter use [the conjunctive one]… and for the other use, standing in for the noun, they express it through a demonstrative pronoun… so that they say: the wind, it, blows it away.
This turn of phrase, the authors note, carries over into the New Testament — in the passage where the apostle Peter, echoing a line from Isaiah, says of Christ that “by his wound you were healed,” the same Hebraic doubling appears. Grammarians, unable to explain this doubling, dismissed it as a meaningless pleonasm — but once you know that the Hebrew relative retains only its conjunctive function, the authors argue, it is not irrational at all.
The second example sits inside Latin itself. Livy, describing the tribune Marcus Flavius proposing that the people of Tusculum be punished, writes “by their aid and counsel” (quorum eorum ope ac consilio) — placing the demonstrative pronoun eorum (“their”) right after quorum, a relative pronoun that already means “of whom/theirs.” This looked so odd that some scholars assumed quorum was a scribal error for quod (“that”). The authors point out that the best manuscripts and editions all agree on quorum eorum — not an error, they argue, but proof that quorum here functions purely as a conjunction. The same construction (quorum eorum unus…) turns up, they note, in Plautus’s comedy Trinummus.
The same principle, the authors continue, resolves a famous dispute among grammarians over the Latin quod — as when Cicero says “I am not reproaching you for the fact that you robbed that man” (Non tibi objicio quod hominem spoliasti). One camp calls this quod an adverb or conjunction; another calls it the neuter form of the relative pronoun itself. The authors’ answer is that both are half right: quod remains a relative pronoun, still bearing a relation to its antecedent, but it has shed its pronominal content (no longer forming part of the subject or attribute) and keeps only its conjunctive function — demoting the complete proposition hominem spoliasti to a mere clause within the main sentence, Non tibi objicio.
And the same principle, they add, applies directly to the French que. “I suppose that you will be wise” (je suppose que vous serez sage) uses a que that is the direct descendant of Latin quod, having shed its pronominal nature in exactly the same way and left only a mark of conjunction.
4. Type 3 — Losing Conjunction Force, Keeping Only the Pronoun
The reverse of Type 2 also occurs: a relative pronoun can lose nearly all its conjunctive force and keep only its demonstrative role. The authors again give two examples.
The first is a stylistic habit peculiar to Latin. Latin writers often open a sentence with a relative pronoun — something untranslatable into the vernaculars, which must substitute a demonstrative instead. Pliny the Younger’s Panegyricus (in praise of the emperor Trajan) opens this way: “Well and wisely did our ancestors establish that the beginning of both action and speech should be taken from prayer… by which custom (Qui mos)… could there be anyone more fitting than a consul to observe and honor it?” This Qui (a relative pronoun) merely opens a new sentence picking up from the one before — the authors note that translating it into French should use “cette coutume” (this custom), not “laquelle coutume” (which custom). The relative here has all but lost its connective force and become a plain demonstrative. Cicero’s Against Verres, they add, is full of similar cases.
The second example is the Greek ὅτι (hoti). This word often corresponds to Latin quod and derives from the same relative pronoun, yet the authors point out that its behavior runs in the opposite direction. Where quod keeps its conjunctive force and sheds its pronominal content, ὅτι often does the reverse — losing its conjunctive force and keeping only its pronominal content. In Revelation chapter 3, where Christ rebukes a self-satisfied church (the bishop of Laodicea), he says: “you say, ‘I am rich’” (λέγεις ὅτι πλούσιος εἰμι). Here ὅτι does not mean “that I [the one speaking to you] am rich” — it instead introduces two separate, independent statements: “you say this: ‘I am rich.’” ὅτι functions here neither as relative nor as conjunction — what remains, the authors suggest, is simply a mark equivalent to “namely.”
Summary
| Type | What the relative retains | Examples |
|---|---|---|
| 1. Pronoun + conjunction (standard) | Both roles intact | Latin in quibus (Livy, on Junius Brutus) |
| 2. Only the conjunction | Loses pronominal content, keeps only a link between two clauses | Hebrew quem…eum (paralleled at 1 Peter 2:24), Latin quorum eorum (Livy, Plautus), quod (Cicero), French que |
| 3. Only the pronoun | Loses connective force, keeps only demonstrative pointing | Latin sentence-initial qui (Pliny’s Panegyricus, Cicero’s Against Verres), Greek ὅτι (Revelation 3:17) |
Why It Still Matters
Modern syntax handles this same duality through terms like “head-internal relative clauses” and “complementizers,” and the double life of French que or English that — sometimes a relative pronoun, sometimes a subordinate-clause marker — is a textbook opening move in syntactic theory. But Arnauld and Lancelot arrived at that distinction not as an abstract rule but by cross-checking real sentences across Latin, Hebrew, Greek, and French — scrutinizing details down to biblical phrasing and manuscript variants along the way. Their discovery that a single word can function, in different cases, as pronoun, as conjunction, or as both at once stands as one of the most precise answers they ever gave to the question they kept returning to throughout this grammar: how many judgments can hide inside a single sentence?