The 100 verbs that carry spoken Finnish

The top line is not close. Even after excluding the negative verb, olla appears more than thirteen times as often as tulla: 22,833 against 1,719.

We searched every Arkisyn token whose part-of-speech tag begins v:, grouped the results by lemma and ranked the counts.¹² That folds oon, oot, on, oli and olis into the dictionary label olla; the descriptive grammar documents why one Finnish verb supplies so many person, tense and mood forms.³

The top 20: the verbs carrying the room

The gap between rank 1 and rank 3 is twelve times the gap between rank 3 and rank 100. The top of this list is therefore not a ranking so much as one verb, one negator, and then everything else. A lemma is also a filing label: the dictionary form is often the form you will never hear, and what arrives in the room is an inflected, contracted version of it. That is what the right-hand column shows.

Pitää alone can hold an object, express obligation or mean liking something, depending on its frame. A learner who knows only “hold” knows its least useful conversational life.

Number two is an entire negative verb

Finnish grammars classify ei as a verb because it inflects for person.⁴

Discarding ei as a grammatical nuisance would erase 13.5% of the 62,913 verb-tagged hits returned by this query. For a listener, en, et and ei are not decoration: they reverse the clause.

21-100: the full count

The ordering inside this table is much softer than the ordering above it. It opens at 297 and closes at 47, so rank 90 and rank 100 are ten places and seven occurrences apart. A different set of recorded conversations would reshuffle that stretch and leave the top twenty exactly where it is. Read this half as a set of verbs to know, not as a league table.

What it does show is how quickly Finnish conversation stops needing new verbs. Ninety-nine lemmas after olla, the counts are already low enough that a single talkative participant could have supplied them on their own.

Equal counts are alphabetised. That is why aloittaa and elää occupy the last two displayed slots while löytyä and syntyä, also on 47, fall outside. The “top 100” has a hard edge the data does not.

Five verbs that disguise themselves in the room

Katsoa → kattoo / kato / katotaan. Katotaan can mean “we’ll see”, including the socially useful version that schedules nothing.

Tietää → tiiän / tiiä. En mä tiiä may express uncertainty or buy a moment while the speaker thinks.

Lähteä → lähtee / lähen. Mä lähen is “I’m leaving”; Lähetäänkö? is “shall we go?” The d in standard lähden and lähdetään disappears in these common southern forms.

Laittaa. It puts dinner on, sends a message, switches the sauna heater on and makes coffee. Translate the scene, not the lemma.

Jaksaa. Dictionaries offer “have the strength”. At 22:40, Mä en jaksa often means “I don’t have it in me” or “I can’t be bothered”, not a medical report.

Method, so the list can be challenged

The reproducible query is Korp command=count, corpus ARKISYN, CQP [pos="v:.*"], grouped by lemma. We sorted absolute counts descending and ties alphabetically.² The query matches the two clean verb tags, v:o and v:neg. Eight tokens with mixed tags such as adv v:neg are excluded, so ei is 8,484 rather than all 8,492 tokens whose annotation contains v:neg.

The University of Turku describes the resource as 29 hours and 26 texts of everyday conversation, with lemma, part-of-speech, morphology and syntax annotation.⁵ Korp’s live Size includes 326,946 tokens of every kind. Its guide warns that linguistic annotation can contain errors.⁶

A larger resource, Lahjoita puhetta, gathered roughly 3,600 hours from more than 20,000 people across Finland, but scale alone does not produce a comparable hand-checkable lemma table.⁷ This ranking is therefore strongest at the top. Start by recognising the spoken forms beside the top 20. Treat the lower ranks as a lookup list: this 29-hour sample cannot make rank 74 meaningfully more important than rank 75.