Finnish for busy people

Old Words versus New Words Ending in -i

There are many words that end in -i in Finnish. Unfortunately – due to the many different changes – their inflection in the different cases can cause some confusion and frustration among learners of Finnish. Let’s take a look!

The biggest portion of words ending in -i are new (eg. lasi : lasin : lasia : laseja). However, there are several categories of old words, like vesi (veden, vettä, vesiä), joki (joen, jokea, jokia) or lohi (lohen, lohta, lohia). There are some other unexpected things going on with old words versus new words ending in -i. Learn more now!

I will be listing the following cases as examples for the inflection of these words:

  1. The partitive case
  2. The genetive case
  3. The inessive case
  4. The plural partitive

1. New Words ending in -i

We will start our article on old words versus new words with the new ones. New words are often loanwords. Usually they’re recognisable because they resemble words in other languages, like pankki for ”bank”, or paperi for ”paper”. Loanwords are easier than Finnish words because they don’t undergo as many changes when you add endings: you basically just add the ending to the basic form of the word in most cases.

1.1. Short new words ending in -i

  • Partitive: basic form +a/ä
  • Genetive: basic form +n
  • Missä: basic form +ssa/ssä
  • Plural Partitive: remove –i and add –eja/ejä
Nominative Partitive Genetive Missä Plural Partitive
bussi bussia bussin bussissa busseja
halli hallia hallin hallissa halleja
hissi hissiä hissin hississä hissejä
kahvi kahvia kahvin kahvissa kahveja
kani kania kanin kanissa kaneja
kassi kassia kassin kassissa kasseja
kasvi kasvia kasvin kasvissa kasveja
kortti korttia kortin kortissa kortteja
koti kotia kodin kodissa koteja
kuitti kuittia kuitin kuitissa kuitteja
kurssi kurssia kurssin kurssissa kursseja
laji lajia lajin lajissa lajeja
laki lakia lain laissa lakeja
lasi lasia lasin lasissa laseja
maali maalia maalin maalissa maaleja
mökki mökkiä mökin mökissä mökkejä
pankki pankkia pankin pankissa maaleja
pappi pappia papin papissa pappeja
passi passia passin passissa passeja
peli peliä pelin pelissä pelejä
pussi pussia pussin pussissa pusseja
siili siiliä siilin siilissä siilejä
taksi taksia taksin taksissa takseja
teksti tekstiä tekstin tekstissä tekstejä
tiimi tiimiä tiimin tiimissä tiimejä
tunti tuntia tunnin tunnissa tunteja
tuoli tuolia tuolin tuolissa tuoleja
väri väriä värin värissä värejä

1.2. Long new words ending in -i: group one

As far as I can think, all the long words that end in -i are loanwords. Most of them end in -li or -ri.

While all long words mainly follow the same rules for each case, the plural partitive is an exception. I’m only listing one option for the plural partitive, but it’s good to know that there are often multiple options for this case.

For group one, all words have a short vowel in the one-but-last syllable (lää--ri). This is why these words have the ending –eita/eitä for the plural partitive.

  • Partitive: basic form +a/ä
  • Genetive: basic form +n
  • Missä: basic form +ssa/ssä
  • Plural Partitive for long words: learn more in this article.
Nominative Partitive Genetive Missä Plural Partitive
enkeli enkeliä enkelin enkelissä enkeleitä
kaakeli kaakelia kaakelin kaakelissa kaakeleita
lääkäri lääkäriä lääkärin lääkärissä lääkäreitä
moottori moottoria moottorin moottorissa moottoreita
mutteri mutteria mutterin mutterissa muttereita
naapuri naapuria naapurin naapurissa naapureita
paperi paperia paperin paperissa papereita
seteli seteliä setelin setelissä seteleitä
sipuli sipulia sipulin sipulissa sipuleita
symboli symbolia symbolin symbolissa symboleita

1.3. Long new words ending in -i: group two

For group two, all words have a long vowel in the one-but-last syllable (in-si-nöö-ri). These words have -eja/ejä in the plural partitive.

  • Partitive: basic form +a/ä
  • Genetive: basic form +n
  • Missä: basic form +ssa/ssä
  • Plural Partitive for long words: learn more in this article.
Nominative Partitive Genetive Missä Plural Partitive
amatööri amatööriä amatöörin amatöörissä amatöörejä
appelsiini appelsiinia appelsiinin appelsiinissa appelsiineja
banaani banaania banaanin banaanissa banaaneja
insinööri insinööriä insinöörin insinöörissä insinöörejä
jonglööri jonglööriä jonglöörin jonglöörissä jonglöörejä
krokotiili krokotiilia krokotiilin krokotiilissa krokotiileja
likööri likööriä liköörin liköörissä liköörejä
miljonääri miljonääriä miljonäärin miljonäärissä miljonäärejä

1.4. Long new words ending in -i: group three

Group three is for long new words ending in something else besides -li or -ri (eg. -tti or -si).

  • Partitive: basic form +a/ä
  • Genetive: basic form +n
  • Missä: basic form +ssa/ssä
  • Plural Partitive for long words: –eja/ejä
Nominative Partitive Genetive Missä Plural Partitive
arkkitehti arkkitehtia arkkitehdin arkkitehdissa arkkitehteja
konteksti kontekstia kontekstin kontekstissa konteksteja
poliisi poliisia poliisin poliisissa poliiseja
presidentti presidenttiä presidentin presidentissä presidenttejä
salaatti salaattia salaatin salaatissa salaatteja
turisti turistia turistin turistissa turisteja

2. Old words ending in -i

We’ve come to part two of the old words versus new words battle! Old words are very often nature words. After all, nature has been around for so long that Finns have had names for them since the very beginning. Some words’ age can be confusing, for example äiti “mother” is actually a new Finnish word, even though mothers have been around since the beginning of time!

2.1. The base rule for old words ending in -i

  • Partitive: remove –i and add –ea/-eä
  • Genetive: remove –i and add –en
  • Missä: remove –i and add –essa/-essä
  • Plural partitive: basic form +a/ä (rule for all old i-words)
Nominative Partitive Genetive Missä Plural Partitive
arki arkea arjen arjessa arkia
arpi arpea arven arvessa arpia
happi happea hapen hapessa happia
hauki haukea hauen hauessa haukia
helmi helmeä helmen helmessä helmiä
henki henkeä hengen hengessä henkiä
hirvi hirveä hirven hirvessä hirviä
joki jokea joen joessa jokia
järvi järveä järven järvessä järviä
kivi kiveä kiven kivessä kiviä
kylki kylkeä kyljen kyljessä kylkiä
lahti lahtea lahden lahdessa lahtia
lehti lehteä lehden lehdessä lehtiä
mäki mäkeä en essä mäkiä
nimi nimeä nimen nimessä nimiä
ovi ovea oven ovessa ovia
pilvi pilveä pilven pilvessä pilviä
polvi polvea polven polvessa polvia
putki putkea putken putkessa putkia
ripsi ripseä ripsen ripsessä ripsiä
rupi rupea ruven ruvessa rupia
sormi sormea sormen sormessa sormia
suomi suomea suomen suomessa suomia
talvi talvea talven talvessa talvia
tammi tammea tammen tammessa tammia
tuki tukea tuen tuessa tukia
tähti tähteä tähden tähdessä tähtiä
väki väkeä en essä väkiä

2.2. Old words ending in -si

More old words, but this time with –si at their end. It’s also important that this rule is only for old words, which means new words like kurssi (kurssia) and bussi (bussia) are excluded from this rule (see 1.1.).

  • Partitive: remove –si and add –tta/-ttä
  • Genetive: remove –si and add –den
  • Missä: remove –si and add –dessa/-dessä
  • Plural partitive: basic form +a/ä (rule for all old i-words)
Nominative Partitive Genetive Missä Plural Partitive
kausi kautta kauden kaudessa kausia
kuukausi kuukautta kuukauden kuukaudessa kuukausia
kuusi (6) kuutta kuuden kuudessa kuusia
käsi ttä den dessä käsiä
liesi liettä lieden liedessä liesiä
mesi mettä meden medessä mesiä
susi sutta suden sudessa susia
tosi totta toden todessa tosia
täysi täyttä täyden täydessä täysiä
uusi uutta uuden uudessa uusia
viisi viittä viiden viidessä viisiä

2.3. Old words ending in -li, -ni or -ri

This rule for words of two syllables ending in –li, -ni or –ri is not 100 % foolproof. There are words that end in –hi, like ”lohi” for example, that become ”lohta” in the partitive.

  • Partitive: remove –i and add –ta/-tä
  • Genetive: remove –i and add –en
  • Missä: remove –i and add –essa/-essä
  • Plural partitive: basic form +a/ä (rule for all old i-words)
Nominative Partitive Genetive Missä Plural Partitive
hiili hiil hiilen hiilessä hiiliä
hiiri hiir hiiren hiiressä hiiriä
huuli huulta huulen huulessa huulia
jousi jousta jousen jousessa jousia
juuri juurta juuren juuressa juuria
kieli kiel kielen kielessä kieliä
kuori kuorta kuoren kuoressa kuoria
kusi kusta kusen kusessa kusia
kuusi (spruce) kuusta kuusen kuusessa kuusia
lohi lohta lohen lohessa lohia
pieni pien pienen pienessä pieniä
saari saarta saaren saaressa saaria
sieni sien sienen sienessä sieniä
suoni suonta suonen suonessa suonia
suuri suurta suuren suuressa suuria
tuli tulta tulen tulessa tulia
tuuli tuulta tuulen tuulessa tuulia
vuori vuorta vuoren vuoressa vuoria
ääni ään äänen äänessä ääniä

2.4. Old words ending in -mi or -hi with two stems

Some words (eg. toimi) have two stems: a vowel stem (toime-a) and a consonant stem (toin-ta) for the singular partitive. Both of these forms are correct, but usually one of them is more common than the other.

The numbers mentioned are the number of results you get when googling these forms. This is not a 100% accurate representation of the reality, but gives you some idea nevertheless.

  • Partitive: either –ea/eä or –ta/tä (-nta/-ntä for -mi words)
  • Genetive: remove –i and add –en
  • Missä: remove –i and add –essa/-essä
  • Plural partitive: basic form +a/ä (rule for all old i-words)
Nominative Partitive #1 Partitive #2 Genetive Missä Plural Partitive
liemi niemeä – 1 300
lien– 138 000
liemen liemessä liemiä
luomi luomea – 9 000
luonta – 17 000
luoen luomessa luomia
riihi riiheä – 500
riih– 7 000
riihen riihessä riihiä
veitsi veitseä – 10 500
veis– 240 000
veitsen veitsessä veitsiä
vuohi vuohea– 1 800
vuohta – 28 000
vuohen vuohessa vuohiä
loimi loimea – 23 000
lointa – 27 000
loimen loimessa loimia
toimi toimea – 38 000
tointa – 62 000
toimen toimessa toimia
niemi niemeä – 31 000
nien– 4 000
niemen niemessä niemiä
taimi taimea – 85 000
tainta – 120 000
taimen taimessa taimia
tuomi tuomea – 11 000
tuonta – 3 000
tuomen tuomessa tuomia

In addition to the words above, it’s also interesting to mention kaali and viini. In some forms of Finnish slang, these words have an exceptional partitive form. Insteal of the regular kaalia and viiniä, some people use the alternative forms kaalta and viin. These are not considered correct, but they are interesting, because the exceptional form follows the regular rule for words ending in –li and –ni.

2.5. Exceptional words

The following words all have some kind of exceptional change going on. The exceptional thing has been marked in purple. For these words the battle is not between old versus new; they’re just exceptional forms, usually due to historical reasons.

Nominative Partitive Genetive Missä Plural Partitive
lapsi lasta lapsen lapsessa lapsia
yksi yh yhden yhdessä yksiä
kaksi kahta kahden kahdessä kaksia
varsi vartta varren varressa varsia
hirsi hirttä hirren hirressä hirsiä
kansi kantta kannen kanness kansia
meri merta meren meressä meriä
veri verta veren veressä veriä
0 0 vote
Article Rating
Notify of

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Newest Most Voted
Inline Feedbacks
View all comments
Michael Hämäläinen

This is very interesting work, thanks for breaking the topic down into manageable chunks!
One inspiration for me to learn to separately analyse consonant clusters depending on where they appear in the word (initial/medial/final) was the Phonotactics section of the Finnish Phonology Wikipedia page. Some of the principles introduced there not only give some hints about the final consonant clusters (e.g., Finnish words only end in the consonants t,s,n,r, or l; all others will be given an –i– ending), but also the initial –s– dropping from some foreign loanwords: e.g., Swedish glas = Finnish lasi; Old Swedish stol = Finnish tuoli (namely, because only stop+liquid combinations are allowed for word-initial consonant clusters).
For those who are interested in theoretical frameworks, inflection stems are in fact the basis for the KOTUS nominal classification system, as explained in the Finnish: Nominal Inflection appendix page on Wiktionary. As that classification system shows, the wide variation in –i– ending nominative forms is contrastive with the consistency and simplicity of nominals ending in round vowels (u, o, y, ö).
Specific to the topic of nominative forms ending in –i-, these are the corresponding KOTUS types:
1.1. Short new words ending in –i
KOTUS type: all 5 risti
Note that 5 risti types can undergo consonant gradation (e.g., koti:kodin). This type includes both two-syllable nominals and words formed by adding the suffix -kki. The 5 risti declension patterns are also applied to foreign words ending in a consonant (e.g. kalsium calcium, Robert (name)).
1.2. Long new words ending in –i: group one
KOTUS types: all 6 paperi
In contrast to 5 risti types, these words do not undergo consonant gradation.
1.3. Long new words ending in –i: group two
KOTUS types: banaani and likööri are 6 paperi; all others are 5 risti
If I understand correctly, banaani and likööri (which do not have consonant gradation) are uniquely classified as 6 paperi because the can take the alternate plural partitive forms ending in –tA: banaaneita (in addition to banaaneja) and likööreitä (in addition to liköörejä)
1.4. Long new words ending in –i: group three
KOTUS types: poliisi is 6 paperi; all others are 5 risti
Note that in partitive plural, poliiseja is more common than poliiseita.
2.1. The base rule for old words ending in –i
KOTUS type: 7 ovi
Note that äiti (which has Proto-Germanic roots) is KOTUS type 5 risti.
2.2. Old words ending in –si
KOTUS types: all 27 käsi
2.3. Old words ending in –li, –ni or –ri
KOTUS types: lohi and tuli are 23 tiili; hiili, hiiri, huuli, kusi, kuusi (spruce) are 24 uni; all others are 26 pieni
There are not many Finnish words ending in –hi, but they show wide variation in declension types, including: 5 risti (kaihi, karhi, orhi, pihi, vihi); 7 ovi (hanhi, kärhi, närhi, tilhi); 23 tiili (jouhi, riihi, tuohi, uuhi); 24 uni (ruuhi)
KOTUS 26 pieni-type are two-syllable nominals ending with –li/-ri/-ni that don’t have consonant gradation. Their declension is essentially identical to 24 uni, except that –ten genitive plural ending is preferred (-en is also possible). 23 tiili (the single –ni ending word “moni” and two-syllable nominals ending in –li/-hi) also has the same declension patterns, but it only takes genitive plural –en suffix.
2.4. Old words ending in –mi or –hi with two stems
KOTUS types: riihi and vuohi are 23 tiili; veitsi is 30 veitsi; all others are 25 toimi.
30 veitsi is exceptional: it is comprised only of the words veitsi and peitsi. 25 toimi type are all two-syllable nominals ending with –mi.
2.5. Exceptional words
meri is 24 uni; veri is 26 pieni; varsi and hirsi are 28 kynsi; lapsi is 29 lapsi (sole member of this type); yksi and kaksi are 31 kaksi (total 4 members; also: haaksi and oksi).
28 kynsi type are nominals ending with –nsi/-rsi/-lsi in which the inflectional stem replaces –si with –te-.
We can see from the above examples that the presence of consonants n, l, r can often impact the declension patterns, also –h– (though less common). A more detailed breakdown is shown in the Finnish: Nominal Inflection appendix mentioned above.

Michael Hämäläinen

My previous comment addressed the theoretical framework of the KOTUS classification, for which inflection stems are taken as the starting point. However, most language teachers or students will find it too unwieldy for use in the classroom.

However, there is a heterodox school of thought that also takes as starting point the inflection stem of the nominals instead of the (dictionary form) nominative form. The motivation for this approach is the realisation that, ironically, the nominative dictionary form is the most “lossy” (i.e., letters in the inflection stem are dropped) and irregular among the 14+ case forms, which means that often it is more complicated to move from nominative form to inflectional stem form, rather than in the reverse direction.

One such pedagogical approach was used in Daniel M Abondolo’s Colloquial Finnish: The Complete Course for Beginners, which involved creating an artificial ‘mark-up’ language — essentially, extra capital letters and symbols to represent the (strong form) stems of various words. Because the underlying structures are already “encoded”, only a few simple steps are needed to “back out” the nominative or other case forms – usually, applying rules of consonant gradation, or simply dropping or swapping the capital letters. He applies a similar approach to verbs, collapsing them down into only 4 classes instead of the standard 6 verbtypes, as I touched upon in a separate comment.

An interesting synthesis of the standard nominative-based categorisation scheme and Abondolo’s mark-up approach is provided in Heinrich Tsanov’s Complete guide to all Finnish noun declensions (almost). The article begins with a self-described “excessively detailed” decision tree starting from the nominative form and leading to the inflection stem type. The main text of the article explains phonological changes and the etymological reasons for splits in the decision tree, referencing many of the principles discussed here.

Again, I will emphasise that this is a heterodox approach — although Abondolo is a respected professor of linguistics and his book (published by Routledge) has sold well, his unique approach has not caught on broadly. In short, there are tradeoffs to this alternative “over-analytical” approach, but for those learners who want to make the sacrifices to learn a logically-rigorous system, I can vouch that the underlying principles are sound.

Inge (admin)

Ooh, that decision tree is cool, I like it 🙂

The approach of using the stem of the word rather than the nominative is something that’s been discussed among S2-teachers. It’s a viable approach for those trying to learn the GRAMMAR of Finnish, but not so much for those who want to talk their simple Finnish from the start. Their basic sentences would be filled with forms that only make sense from a grammatical point of view. Theoretic forms like vede- and pöydä- don’t make much sence when you say “vede- ole- pöydä-” (The water is on the table). Which isn’t to say that “vesi olla pöytä” is much clearer either. Try explaining (in Finnish!) to an immigrant that “you’re learning these forms but they’re NOT words, you can use none of these as they are in sentences”. It requires quite a linguistic ability to look at language abstractly enough to grasp the concept of “stems”.

Immigrants from Russia generally know their stems, if they’ve studied Finnish before they moved. Their course books are (if I understand correctly) still from the period where it was common to learn the stems at the same time as the nominative. So they learn vesi : vede- : vete-.

That said, in groups for students with no school background at all and no literacy skills, at least verbs are often taught using the third person singular or the stem. Over time, we’ve come to the conclusion that – if we want to reduce the number of misunderstandings when talking to someone with no ability to grasp the idea of conjugation – it’s clearer to have them remember “opiskele” rather than “opiskelen” or “opiskella”. You will understand “minä opiskele, sinä opiskele, hän opiskele” with less ambiguity than the alternatives: “minä opiskelen, sinä opiskelen, hän opiskelen” or “minä opiskella, sinä opiskella, hän opiskella”. That’s a compromise that has to be made at the beginning of their studies. The reality is that learning to conjugate verbs can be really hard with no background of analyzing language at all. Likewise, they usually just learn “vettä” rather than “vesi” because that’s the form you will need in most cases.

Michael Hämäläinen

Thank you for the insight into the pedagogical side.

Abondolo’s book makes no reference to why he created his unique approach, and I remember asking an experienced Finnish teacher in London about strange words like “huoneQ” (“huone” with glottal stop / boundary gemination marker marked with capital Q” and “vieraX” (the X becomes ‘s‘ in nominative, “vieras“). She had no idea what I was talking about. So even though Abondolo’s book is probably within the top 10 most popular English-language texts, his approach hasn’t caught on at all. Perhaps as a pedagogical method, it is a non-starter, unless an entire infrastructure (including textbooks and reference materials) were built up around it.

The point about Russian textbooks is quite interesting. Perhaps for speakers of languages with many cases and complicated inflections, it is easier to “bite the bullet” and lay the inflection stem groundwork, no matter if the initial investment is onerous. If I were to make an analogy, mainland Chinese students start their study of English by first learning the International Phonetic Alphabet (IPA), whereas lots of Japanese schoolchildren take the shortcut of marking English pronunciation with katakana (Japanese phonetic writing system). I myself only learned the value of the IPA when I took some intro courses on Swedish (complicated vowel system!) and Russian (palatisation!)

So, as a self-study methodology, I see promise in these heterodox approaches. Wiktionary is a game-changer for me, as it provides etymologies, derivation forms, KOTUS type groupings, and complete inflection tables for nearly every word — by contrast, textbook explanations leave me cold because I know they have been pared down and simplified, but I don’t know what was left on the cutting room floor. The decision trees for classifying the words are complicated, but I just make a bunch of blank templates on the office copy machine and practice filling them out once a day, just like a crossword in the morning newspaper.


I’m quite new to this site and I am so happy to see this sort of additional info being provided by you. Your comments over these articles have been very helpful in bringing order and structure in the progress to fluency.

I have actually worked through the Abondolo book, and I must say I found his boundary gemination marker system (huoneQ, vieraX, etc) confusing rather than helpful.

Another writer (I don’t recall which one) used a simplified system of just marking boundary gemination with a superscript X, which was better, because the phenomenon is only relevant to the spoken language, and the extra or doubled sound always takes on the initial consonant of the next word.

BTW, I love your extracts from the Korpela material. I have his handbook (the giant version of 900+ pages) and have been wading my way through it, on and off, for a couple of months. It’s always good to get another take on his very extensive and useful material.