Someone asks about your weekend. You start talking. El sábado… and then the fork: fui or iba? Estuve or estaba? Hizo or hacía?
You know the rules. Completed action, indefinido. Ongoing or background, imperfecto. You have filled in the worksheets and passed the tests. But mid-sentence, with someone waiting, the rule does not fire fast enough. So you pick one, usually the indefinido because it feels more finished, and carry on. Nobody corrects you.
You made the safe choice again, and there is a name for the pattern.
The short answer
You are not confusing two tenses. You have one default past tense and you are applying it to everything, which is the documented early behaviour of English-speaking learners of Spanish. Rafael Salaberry analysed oral narratives from twenty college-level learners and found the preterite functioning as a general past marker, used independently of the inherent aspect of the verb.
And the rule you were taught cannot rescue you, because it describes the wrong thing. The choice is not completed against ongoing. It is which frame you are putting the event in, and framing is a decision you make before you choose a tense, in about half a second, while building the rest of the sentence.
The rules work on paper, and speech is not paper
The textbook version: the pretérito indefinido describes completed actions, the pretérito imperfecto describes ongoing states, habits and background. The clean pair is ayer llovió against ayer llovía cuando salí.
Useful as far as it goes. Then real conversation hands you el hotel era bonito pero la habitación no tenía aire acondicionado, two imperfects describing what sound like completed observations. Or viví en Madrid tres años, an indefinido for something that sounds thoroughly ongoing. The rule starts leaking immediately.
It is not wrong so much as pitched at the wrong level. Both tenses encode grammatical aspect, which is not about when something happened but about how the speaker presents its internal shape: as a bounded whole, or as something with an inside you can put other events into.
The same reality, two framings
Take one event: you lived in Barcelona for two years.
Viví en Barcelona dos años. Indefinido. A closed chapter, a fact in your biography, complete.
Vivía en Barcelona cuando conocí a Ana. Imperfecto. A backdrop, an open scene with room in it for something else to happen.
Same reality, both correct, different jobs. The choice depends on what the sentence is for, and that is a perspective decision rather than a fact about the world.
Your default has a research literature
The intuition that you lean on the indefinido is correct, and it is not a personal quirk.
Salaberry’s 1999 study collected oral movie narratives from twenty English-speaking learners at two points two months apart, with native speakers as a control, and found the preterite serving as a default past marker in the early stages, applied regardless of the verb’s inherent aspectual value. The prediction that grew out of this is usually called the Default Past Tense Hypothesis, and it says the association between aspect and tense marking gets stronger as proficiency rises rather than being there from the start.
The wider frame is the Aspect Hypothesis, reviewed across two decades of studies by Kathleen Bardovi-Harlig and Llorenç Comajoan-Colomé in Studies in Second Language Acquisition. The hypothesis predicts learners first attach past marking to perfective verbs before extending it, and their honest conclusion is that the picture is messier than the original theory: lexical aspect, discourse context, first language and individual variation all interact.
The useful takeaway from the mess is the direction. Early learners have one past tense that does everything. Getting two is a developmental step, not a matter of remembering harder.
It is worth sitting with that for a second, because it reframes the frustration. If you have been treating this as a memory failure, you have been trying to fix something that was never broken. You remember the rule perfectly. What you do not yet have is a second gear, and gears are built rather than recalled. The same reframing applies across understanding far more than you can produce.
Why the default is invisible
Fui al médico instead of iba al médico cuando me llamaste. Fue una buena película instead of era una historia sobre. You reach for the closed, definite version of everything, and since it is not technically wrong, nobody says anything.
This is avoidance in the technical sense. Elaine Schachter’s 1974 paper found learners producing almost no relative clauses rather than producing them badly, which made an error count rank them as strong. Martin Kleinmann’s follow-up identified the trigger as confidence in that one structure. Both describe your imperfect exactly, and the mood version of the same dodge is subjunctive avoidance.
What natives notice is not an error but a texture. Your stories do not breathe. Everything is sequential, fui, comí, volví, with no background, no scene, no shift of perspective. Grammatically clean and conversationally monotone.
The social cost of that is easy to underestimate, because it does not look like a language problem from the outside. It looks like you not having much to say. A story without a background layer gives the listener nowhere to react until it ends, so the small interjections that make Spanish conversation feel collaborative never get a foothold, and you end up doing monologues where everyone else is doing exchanges. The repair kit for that is partly discourse markers and partly this: the imperfect is what gives a listener somewhere to stand.
Why more rule study does not help
Ask any intermediate learner to explain the difference and you will get a reasonable answer: completed against ongoing, snapshot against video, the keyword lists with siempre and todos los días on one side and ayer and de repente on the other.
None of it fires in time, for three reasons.
The keywords are usually absent. Real sentences do not arrive with ayer attached. You choose from your intended framing, and trigger-hunting is the one skill the exercises did train.
Both options are often correct. Conocí a Juan en la fiesta and conocía a Juan de la universidad are both fine and mean different things. This is not right against wrong, it is which meaning you want, decided live.
Rules run slower than speech. Even a correctly firing rule takes two or three seconds. Conversation offers about half of one. Norman Segalowitz’s three kinds of fluency name the binding constraint as cognitive fluency, the speed of the machinery itself, and no amount of additional declarative knowledge speeds it up. That is the same wall behind conjugation tables not making you conjugate.
Robert DeKeyser’s skill acquisition account explains why the knowledge is not wasted and is also not enough: declarative knowledge is a real stage, and it becomes usable only through practice under the conditions of use. You have completed the first stage several times.
The strongest case for the default
Before the fix, the honest counter-argument.
A default is efficient. One past tense that covers everything is a working system, it is understood every time, and the cost of the missing nuance is paid by the listener in a currency they rarely notice. For a learner whose goal is to be understood, spending months on a distinction that changes texture rather than meaning is a poor trade against, say, vocabulary.
The developmental evidence also suggests patience is not laziness. If the association between aspect and tense marking strengthens with proficiency, as Salaberry argues, then some of this arrives on its own as the rest of your Spanish grows. Forcing the imperfect early can produce a learner who sprinkles era and tenía into places they do not belong, which is a worse outcome than a clean, flat narrative.
Why it still has to be trained
The catch is that “arrives with proficiency” describes learners who keep producing narrative. Most adults stop.
The default is self-reinforcing, because every story you tell in one gear is a rep for the one gear. Merrill Swain’s output hypothesis points at what is missing: comprehension lets you succeed without processing the form, and only production forces the choice. But production that always takes the same exit forces nothing. You can narrate for a decade and practise the default a decade’s worth.
What builds the production skill
Narrate, do not conjugate. Tell real stories about your day, your weekend, your childhood. Force a scene before the events: hacía frío, la calle estaba vacía, then entré en el bar. The imperfect becomes natural when it is used to paint a backdrop rather than to fill a blank.
Retell the same story two ways. Once as a summary, mostly indefinido: fui, vi, comí, volví. Then as a scene, mostly imperfecto: hacía calor, la gente paseaba, olía a mar. Switching framing on identical content is the drill, and it exists in no worksheet.
Retrieve rather than review. Roediger and Karpicke’s test-enhanced learning work found restudying winning on an immediate test and losing badly on a delayed one. Re-reading the rule is restudying. Producing a narrative is the test.
Spread it out. Cepeda and colleagues synthesised 839 assessments of distributed practice across 317 experiments and found spacing reliably beating massed repetition. Four short narrations across four days beat one long session, and will feel less productive throughout.
Keep it a portion of the diet. Paul Nation’s four strands put deliberate study at roughly a quarter alongside input, output and fluency work. Aspect drills alone produce someone who can explain the distinction beautifully.
Read narrative, not just anything. The background layer is far denser in fiction and in people telling stories than in news or textbooks, so what you read changes what you absorb. This is the vocabulary point applied to grammar: items move from recognition toward use only when you meet them somewhere you might plausibly produce them, which is the ladder we set out in passive vs active vocabulary.
What we do with this at Suelto
The past tenses are treated as a production target rather than a grammar point, because the evidence says the knowledge is already there.
Production tasks are storytelling situations: describe what happened, set the scene, explain the background. The system tracks whether you reach for both tenses or collapse everything into one, so a narrative of pure indefinido registers as a signal rather than as a correct answer. It is correct, and that is exactly why it needs recording.
Territory keeps the two tenses as separate states rather than one past-tense category, because our own data shows learners holding one solidly and the other not at all. The verb step is timed, since selection speed is what decides whether the imperfect is available when the sentence is already moving.
The gap test shows whether your past-tense Spanish is building scenes or listing facts, in about ten minutes with no signup. The method page sets out what the rest rests on.
You do not have a rules problem. You have one gear, the research expected exactly that, and the second one gets built by telling the same story twice rather than by reading the explanation one more time.