Twenty passato prossimo exercises in a row feel wonderful. By number seven you have the pattern; by number twenty you are flying, and you close the book convinced. Then Thursday’s conversation requires you to pick between passato prossimo and imperfetto in real time, and the fluency is gone, because picking the tense was the skill, and your practice never once asked you to pick.

This is the difference between blocked practice (AAAA BBBB CCCC) and interleaved practice (ABCACB…), and the research verdict is one of the most consistent in learning science. Rohrer and Taylor’s mathematics studies are the famous case: students practising mixed problem types performed worse during practice than blockers, and dramatically better on the later test. The pattern repeats across domains, and it belongs to the family Robert Bjork named desirable difficulties: conditions that impair practice performance while improving learning, the same family as spacing and testing.

Why mixing teaches choosing

Blocked practice quietly removes the hardest part of the task. When every exercise is the passato prossimo, you never ask “which tense does this sentence need?”; the block answered it for you. But discrimination is the skill: real Italian arrives unlabelled. Interleaving forces every item through the full decision (identify the situation, select the structure, apply it), which is exactly the pipeline conversation uses. It also spaces things automatically: mixing three tenses means each one keeps coming back after a gap, so you get the spacing effect for free.

The cost is honest: interleaved practice feels worse. More errors, less flow, no moment where you are flying. Learners reliably judge it inferior, in Italian as in the lab. Trust the Thursday test, not the Tuesday feeling.

Interleaving: Why Mixed Italian Practice Works Better

Interleaving your Italian

  • Mix contrasting grammar deliberately: the pairs you confuse are the pairs to shuffle. Passato prossimo with imperfetto. Essere with stare. Per with da. A mixed set of fifteen beats three pure sets of twenty.
  • Make reviews mixed by default: your error queue is naturally interleaved: today’s review might hold a preposition, a plural and a congiuntivo. That mess is a feature.
  • Block briefly, then mix: the research does not forbid an initial block to meet a new structure. Learn the form for ten minutes; then it goes into the mix and stays there.
  • Revisit old grammar inside new topics: which is the spiral idea from grammar in texts, interleaving at the scale of a whole course.

How amo Italian mixes

In amo Italian, interleaving is policy at three levels. Within a tappa, twenty-plus exercise types rotate, so the format never lets you coast on autopilot. The Facciamo il punto review stages that punctuate every level re-test earlier grammar mixed together in new contexts, deliberately never introducing anything new, only reshuffling what came before. And the spaced review queue draws from your whole history of real errors across all topics, which makes every session a mixed set by construction.

The summary is uncomfortable and liberating: if practice feels smooth, it is probably teaching you less than it feels like. Shuffle the deck.

References and attributions: Rohrer & Taylor, “The shuffling of mathematics problems improves learning”, Instructional Science, 2007; Bjork on desirable difficulties. “Desirable difficulties” is Robert Bjork’s term. Findings are paraphrased with attribution. Part of the Science of Learning Italian series.