Academic Catalan often relates a claim to its evidence, limitations, conditions, and source. The challenge is to build a hierarchy in which every clause has a recoverable function, every modal has the intended scope, and every pronoun has one credible antecedent.
English academic prose can stack nouns and repeat overt subjects. Catalan more readily uses subordinate clauses, null subjects, and weak pronouns. Those resources create elegant compression, but only when the writer controls the links they leave implicit.
Think in levels, not chains
A layered sentence has a matrix clause and one or more subordinate domains. Connectors at different levels do not form a flat string.
L'estudi conclou que, si es mantenen constants les variables que afecten la demanda, l'augment observat no es pot atribuir únicament a la nova tarifa.
The study concludes that, if the variables affecting demand are held constant, the observed increase cannot be attributed solely to the new rate.
Its structure is:
[Matrix L'estudi conclou [content que [condition si es mantenen constants les variables [relative *que afecten la demanda]], [claim l'augment observat no es pot atribuir únicament a la nova tarifa]]].
The si-clause conditions the embedded claim, not the act of concluding. The relative que afecten la demanda modifies only les variables. Punctuation helps processing, but the conjunctions and argument relations create the hierarchy.
| Level | Typical job | Useful forms |
|---|---|---|
| matrix | attribute or qualify the claim | l'estudi sosté, les dades suggereixen, no podem afirmar |
| content | state what is claimed | que, si, interrogative clause |
| frame | limit time, condition, cause, or concession | si, com que, tot i que, quan |
| modifier | identify a referent | relative or participial clause |
Label these levels before polishing the wording. If they cannot be bracketed consistently, readers will not recover them consistently.
Mood belongs to each local dependency
There is no “academic subjunctive” applied to a whole sentence. Each embedding environment assigns its own mood according to assertion, modality, and connector meaning.
Els autors admeten que, tot i que les dades indiquen una tendència estable, cal que s'ampliï la mostra abans que se'n generalitzin els resultats.
The authors acknowledge that, although the data indicate a stable trend, the sample must be enlarged before its results are generalized.
Admeten que asserts a content clause, and factual tot i que contains indicative indiquen. Impersonal necessity cal que selects subjunctive s'ampliï. Prospective abans que selects subjunctive se'n generalitzin, where en resumes de la mostra. The sentence contains three subordinate relations, but mood is decided separately at every joint.
Negation also changes scope according to its position:
No afirmem que aquesta correlació demostri una relació causal.
We do not claim that this correlation proves a causal relationship.
Afirmem que aquesta correlació no demostra una relació causal.
We claim that this correlation does not prove a causal relationship.
The first withdraws commitment to the embedded proposition and uses demostri. The second asserts a negative proposition and uses indicative demostra. English displays the scope contrast too, but its verb form does not force the writer to notice it.
Connector hierarchy and information structure
Put the widest frame where the reader can recognize it early. Within that frame, order narrower concessions, conditions, and evidence close to the proposition they qualify.
Atès que les sèries anteriors no són comparables i que, encara que s'hi apliqui la mateixa ponderació, persisteix un biaix de selecció, convé interpretar amb prudència la diferència observada.
Given that the earlier series are not comparable and that, even if the same weighting is applied to them, a selection bias persists, the observed difference should be interpreted cautiously. (academic)
The coordinated causes jointly license the recommendation. The concessive sits inside the second cause: applying the same weighting does not remove the bias. Moving encara que… after the recommendation could make it appear to qualify only interpretar amb prudència.
A hierarchy can also be distributed across sentences. C2 style is not a contest in maximum embedding:
Les sèries anteriors no són comparables. A més, el biaix de selecció persisteix encara que s'hi apliqui la mateixa ponderació; per tant, la diferència observada s'ha d'interpretar amb prudència.
The earlier series are not comparable. Moreover, the selection bias persists even if the same weighting is applied to them; therefore, the observed difference must be interpreted cautiously.
The second version makes the inference explicit with per tant and lets readers retain the premises separately.
Referent tracking: pronouns save words, nouns save readers
Catalan null subjects and clitics efficiently maintain an established referent. They should not force the reader to choose among equally plausible antecedents.
La comissió va trametre l'informe al rectorat perquè el revisés abans de publicar-lo.
The committee sent the report to the rector's office so that it could review it before publishing it.
The direct-object clitic el clearly resumes l'informe, but the null subject of revisés could be the rector's office or, in a poorly established context, another reviewer. Publicar-lo also leaves the publisher implicit. In controlled academic prose, name the actor at the point where it changes:
La comissió va trametre l'informe al rectorat perquè aquest organisme el revisés abans que la comissió el publiqués.
The committee sent the report to the rector's office so that that body could review it before the committee published it.
Repeating la comissió prevents false coreference. (Formal) aquest organisme works because the preceding institution is uniquely recoverable.
After a preposition, the inflected relative el qual/la qual/els quals/les quals can clarify the antecedent in formal writing:
El model incorpora una restricció sense la qual els resultats deixarien de ser comparables.
The model incorporates a constraint without which the results would cease to be comparable. (formal)
Use el qual when its preposition and agreement clarify a relation that bare que would obscure. See restrictive and nonrestrictive relatives.
An ambiguity diagnosis
L'article analitza les interpretacions de les autores que havien estat qüestionades.
The article analyzes the interpretations of the authors that had been challenged.
Feminine plural qüestionades can agree with les interpretacions or les autores. The relative therefore has two syntactically available antecedents. Commas do not magically choose the intended restrictive reading. Rewrite it:
L'article analitza les interpretacions qüestionades que havien proposat aquelles autores.
The article analyzes the disputed interpretations that those authors had proposed.
L'article analitza les interpretacions proposades per les autores que havien estat qüestionades.
The article analyzes the interpretations proposed by the authors who had been challenged.
The first attributes questioning to the interpretations; the second to the authors. Lexical repetition and participial placement make the attachment visible.
Neutral order and a clause used as topic
Neutral academic order puts the source before its content:
Diversos treballs ja advertien que aquesta correlació no implicava causalitat.
Several studies had already warned that this correlation did not imply causation. (neutral order)
The content clause can be fronted as an established topic, but Catalan resumes it with ho:
Que aquesta correlació no implicava causalitat, ja ho advertien diversos treballs.
That this correlation did not imply causation—several studies had already warned of it. (marked topic-first order)
The initial que-clause is not the subject of advertien. The postverbal diversos treballs is the subject, and ho resumes the content object. English analysis based only on the first phrase would miss both relations. The reordered version is useful when the proposition has already been discussed; as discourse-new information, it would sound needlessly theatrical.
Punctuation is a map, not structural glue
Paired commas can enclose a short subordinate frame inside a content clause. A semicolon can separate closely related independent units. Neither mark can replace a missing connector or repair an unattached modifier.
Els resultats suggereixen que, quan augmenta la pressió, el material perd elasticitat; aquesta pèrdua, però, no és irreversible.
The results suggest that, when pressure increases, the material loses elasticity; this loss, however, is not irreversible.
The comma pair brackets the temporal clause; the semicolon opens an independent qualification; commas make però parenthetical.
Nominalization, passives, and responsibility
Dense prose often accumulates nominalizations and agentless passives because English source texts do so. Catalan has both resources, but active or pronominal syntax often reveals the reasoning more directly.
La verificació dels resultats va ser efectuada pels investigadors que havien desenvolupat el mètode.
The verification of the results was carried out by the researchers who had developed the method. (grammatical but heavy)
Els investigadors que havien desenvolupat el mètode van verificar-ne els resultats.
The researchers who had developed the method verified its results. (clear active revision)
The revision makes the agent the subject and uses en for del mètode. A passive still suits a topical patient or an immaterial agent, but formality alone does not justify hiding agency.
A natural supervisory dialogue
—En aquest paràgraf dic que el model falla perquè les dades són incompletes, tot i que després afirmo que la mostra és suficient. —Aleshores has posat dues conclusions incompatibles sota el mateix perquè.
—In this paragraph I say that the model fails because the data are incomplete, although later I claim the sample is sufficient. —Then you have placed two incompatible conclusions under the same perquè.
—Ho separo en dues frases? —Sí: primer limita l'abast de la mostra i després explica en quines condicions el model continua sent útil.
—Should I split it into two sentences? —Yes: first limit the scope of the sample and then explain under which conditions the model remains useful.
Splitting prevents the causal connector from taking unintended scope over the later concession.
A revision workshop
The following sentence is grammatical in fragments but difficult to parse as a whole:
❌ L'informe assenyala que els municipis que van respondre l'enquesta, que es va distribuir quan ja s'havien aprovat els pressupostos, presenten dades que, tot i que són parcials, es considera que permeten conclusions que caldria que fossin verificades.
Poor academic drafting—the stacked relatives, shifting subjects, and remote passive make the evidential status opaque.
Revise by separating source, limitation, and inference:
✅ L'enquesta es va distribuir quan els pressupostos ja s'havien aprovat, i només alguns municipis la van respondre. Per tant, l'informe considera que les dades, encara que parcials, permeten formular conclusions provisionals que caldrà verificar.
The survey was distributed after the budgets had already been approved, and only some municipalities answered it. The report therefore considers that the data, although partial, allow provisional conclusions to be drawn that will need verification.
The revision names the sampling limitation first. Per tant marks the inferential transition, encara que parcials compresses a controlled supplement whose subject is clearly les dades, and future necessity caldrà verificar avoids the awkward nested caldria que fossin verificades.
Regional and register guidance
The principles apply across Catalan varieties. (regional: Valencian) Standard s'amplie, s'aplique, and siguen may replace Central s'ampliï, s'apliqui, and siguin; the hierarchy is unchanged. (formal, IEC/AVL) Atès que, el qual, and per tant are shared resources, but multiplying them does not create precision.
Common Mistakes
1. Letting one mood spread across independent embeddings.
❌ Els autors admeten que les dades indiquin una tendència, però dubten que sigui estable.
Incorrect for the intended assertion—the factual content of admeten takes indicative, while doubt takes subjunctive.
✅ Els autors admeten que les dades indiquen una tendència, però dubten que sigui estable.
The authors acknowledge that the data indicate a trend, but they doubt that it is stable.
2. Omitting a clitic after a fronted content topic.
❌ Que faltaven dades, ja advertia l'annex metodològic.
Incorrect topic construction—the fronted content object must be resumed by ho.
✅ Que faltaven dades, ja ho advertia l'annex metodològic.
That data were missing had already been noted in the methodological appendix.
3. Treating the first phrase as the subject.
❌ Atès que la mostra és petita demostren els resultats una tendència provisional.
Poor information structure—the causal frame needs a comma, and the marked verb–subject order obscures the main subject.
✅ Atès que la mostra és petita, els resultats només apunten a una tendència provisional.
Given that the sample is small, the results only point to a provisional trend.
4. Hiding a relevant agent in a passive stack.
❌ Va ser decidida la retirada de les dades després que fossin detectades errades.
Grammatical fragments but poor controlled prose—the decision-maker and detector disappear.
✅ El comitè va retirar les dades després que l'equip auditor hi va detectar errades.
The committee withdrew the data after the audit team detected errors in them.
5. Attaching a relative to two possible antecedents.
❌ L'estudi compara les respostes de les participants que van ser codificades manualment.
Ambiguous—feminine plural codificades can describe the answers or the participants.
✅ L'estudi compara les respostes que l'equip va codificar manualment amb les de la resta de participants.
The study compares the answers that the team coded manually with those of the remaining participants.
Correction tasks
Scope: Revise No concloem que el tractament no és eficaç if the intended meaning is “we withhold the positive claim that it is effective”:
No concloem que el tractament sigui eficaç.
We do not conclude that the treatment is effective.
Referent tracking: Revise La investigadora va presentar l'autora a la directora abans que parlés:
La investigadora va presentar l'autora a la directora abans que l'autora prengués la paraula.
The researcher introduced the author to the director before the author took the floor.
The repeated noun chooses the subordinate subject. If the director was meant to speak, repeat la directora instead.
Key Takeaways
- Bracket matrix, content, framing, and modifying clauses before polishing the sentence.
- Assign mood locally; no mood applies globally to a multi-clause sentence.
- Place wide-scope connectors early and narrow modifiers beside what they modify.
- Use clitics and null subjects for stable continuity, but repeat a noun when the referent changes.
- Active syntax often makes academic responsibility clearer than a chain of passives.
- Split a sentence when its clause hierarchy no longer mirrors the hierarchy of the argument.
Now practice Catalan
Reading grammar gets you part of the way. The exercises are where it sticks — free, no signup needed.
Start learning Catalan →Related Topics
- Noun Complement ClausesB2 — Build Catalan clauses that supply the content of nouns such as fet, idea, dubte, possibilitat, and esperança.
- Restrictive versus Nonrestrictive RelativesB2 — Use reference, prosody, punctuation, and relative forms to distinguish identifying Catalan relatives from parenthetical ones.
- Epistemic, Deontic, and Dynamic ModalityC1 — How Catalan distinguishes probability, obligation, permission, willingness, and ability through modal verbs and related constructions.
- Advanced Focus ConstructionsC1 — How Catalan uses clefts, pseudoclefts, fronting, focus particles, and prosody to identify or correct the informative center of a sentence.
- Academic StyleB1 — Build Catalan academic arguments through precise definitions, calibrated claims, evidence, impersonal constructions, integrated citations, and coherent paragraphs.