Kwantlen Polytechnic University

KORA: Kwantlen Open Resource Access
All Faculty Scholarship

Faculty Scholarship

2008

Diminutive Verbal Suffixes In Finnish
Päivi Koskinen
Kwantlen Polytechnic University

Solveiga Armoskaite
University of British Columbia

Follow this and additional works at: http://kora.kpu.ca/facultypub
Part of the European Languages and Societies Commons, Modern Languages Commons, and the
Morphology Commons
KORA Citation
Koskinen, Päivi and Armoskaite, Solveiga, "Diminutive Verbal Suffixes In Finnish" (2008). KORA Faculty Scholarship: Paper 31.
http://kora.kpu.ca/facultypub/31

This Conference Paper is brought to you for free and open access by the Faculty Scholarship at KORA: Kwantlen Open Resource Access. It has been
accepted for inclusion in All Faculty Scholarship by an authorized administrator of KORA: Kwantlen Open Resource Access. For more information,
please contact kora@kpu.ca.

DIMINUTIVE VERBAL SUFFIXES IN FINNISH*
Solveiga Armoskaite
University of British Columbia

1.

Päivi Koskinen
Kwantlen Polytechnic University

Introduction

The Finnish suffixes -ahta/-aise are traditionally labeled as aspectual. They are
said to provide the verbal predicate with an interpretation of a momentary action
as (1b,c) illustrate (Hakulinen 1941:46; Karlsson 1983:276; ISK 2005).
(1)

a.
b.
c.

vinku-u
‘whines’
ving-ahta-a ‘gives a brief whine or squeak’
vink-aise-e ‘gives a brief whine or squeak’

Given the descriptive labels aspectual and momentary, we should expect that
either of the two suffixes can attach to all predicates within the same aspectual
class. This is not the case.
(2)

a.

Lapsi nukku -i. 1
child sleep -PAST.3SG
‘The child was sleeping.’

b.

Lapsi nuk -aht -i.
child sleep -ahta -PAST.3SG
‘The child fell asleep.’

c.

* Lapsi el -äht -i.
child live -ahta -PAST.3SG

(2) shows that only some stative predicates can occur with -ahta. This
limited application of the two suffixes holds across all aspectual classes, although
all examples are not included due to lack of space. Furthermore, the contribution
of the suffixes to the interpretation of the predicate varies. In (1b) the acquired
meaning is that of a brief action. In (2b) the acquired meaning is that of change of
state. The lack of consistency in distribution and interpretation makes the
traditional aspectual classification of the suffixes questionable.
* We thank Martina Wiltschko and the UBC Linguistics Thesis Anonymous group
for helpful comments. All errors are our own. Any questions or comments can be
directed to solveiga.armoskaite@gmail.com or paivi.koskinen@kwantlen.ca.
1 In Finnish, the morpho-phonological process of consonant gradation reduces geminate
voiceless stops to single voiceless stops, and single voiceless stops t, p, k to d, v and 0/ (cf.
e.g. Kiparsky 2003: 117, among others). The process of vowel harmony alternates the
front vowels y, ä, ö and back vowels u, a, o in suffixes (cf. e.g. Kiparsky 2003:114).
Actes du congrès annuel de l ’Association canadienne de linguistique 2008.
Proceedings of the 2008 annual conference of the Canadian Linguistic Association.
© 2008 Solveiga Armoskaite and Päivi Koskinen

2
We argue that -ahta/-aise are [+quantized] diminutive modifiers of events,
and that their aspectual effects are only epiphenomenal. Thus, the range of
interpretations that -ahta/-aise contribute is a result of the interaction between the
verbal predicate and the [+quantized] diminutive content of the modifier.
The reminder of the paper is organized as follows. §2 provides evidence
for the modifier status of -ahta/-aise. §3 addresses the aspectual contribution of
the suffixes. §3.1 accounts for their [+quantized] diminutive content and its
interaction with aspectual classes. §3.2 presents further evidence against an
aspectual classification of -ahta/-aise. In §4 we address further questions. §5
concludes the discussion.
2.

Evidence for the Modifier Status of the Suffixes

In this section we present evidence to support our claim that -ahta/-aise are
modifiers. Due to lack of space, we give examples of either -ahta or -aise
throughout the paper, as long as the differentiation between the two is not
relevant for the argument at hand.
First, we discuss the optionality of the two suffixes. Under general
assumptions, modifiers are considered optional, while heads are obligatory.
While the lack of a modifier such as -ahta results in a difference of lexical
meaning, the omission does not produce ungrammaticality, as felicitous examples
(3a) versus (3b) illustrate.
(3)

a.

Lapsi nukku -i.
child sleep -PAST.3SG
‘The child was sleeping.’

b.

Lapsi nuk -aht -i.
child sleep -ahta -PAST.3SG
‘The child fell asleep.’

In contrast, the lack of a head such as Tense results in ungrammaticality, as
in (4a) versus (4b).
(4)

a.

b.

Lapsi nukku -i.
child sleep -PAST.3SG
‘The child was sleeping.’
* Lapsi nukku.
child sleep

3
Second, we note that heads are said to form binary pairs. Such opposing
values may be marked overtly. In English tense marking the third person present
and past inflections differ. This is shown in (5a) versus (5b).
(5)

a.
b.

The boy plays the guitar.
The boy played the guitar.

Alternatively, in binary opposition the presence or absence of a head induces
opposing meanings. In English the first person present lacks overt temporal
inflection while it carries past inflection, as in (6a) versus (6b).
(6)

a.
b.

I play_ the guitar.
I played the guitar.

In contrast, modifiers do not form binary pairs, and the absence of a modifier
does not induce an opposite meaning. -ahta/-aise behave like modifiers. For
instance, while -aise in (7a) induces the reading that the book was read hastily,
the unmarked (7b) does not specify whether the reading was done hastily or not,
i.e. it can mean either if provided with an appropriate context.
(7)

a.

Ignas luk -ais - i
kirjan.
Ignas read -aise -PAST.3SG book.ACC
‘Ignas read a/the book hastily’

b.

Ignas luk -i
kirjan.
Ignas read -PAST.3SG book.ACC
‘Ignas read a/the book.’

A third distinguishing quality between heads and modifiers is the availability
of periphrastic paraphrasing. A modifier allows for a synonymous paraphrase,
while a head does not have such a synonymous option available. Again, -ahta/
-aise act like modifiers: e.g. (7a) can be paraphrased as (8), where the use of -aise
is substituted for by an adverbial modifier.
(8)

Ignas luk -i
kirjan
kymmenessä minuutissa.
Ignas read -PAST.3SG book.ACC ten.INESSIVE minute. INESSIVE
‘Ignas read a/the book in ten minutes’
Interpretation: ‘Ignas read a/the book, but since reading normally takes a
long time and here the reading activity only lasted 10 minutes, he cannot
have paid much attention to what he was reading’

4
In this section we have provided evidence for the modifier status of -aise/
-ahta. We have shown that the suffixes are optional, do not form binary pairs and
allow for periphrastic paraphrasing, none of which are qualities of heads.
3.

The Aspectual Contribution of -ahta/-aise

This section addresses the aspectual contributions of the suffixes. §3.1
establishes the [+quantized] diminutive content and its interaction with aspectual
classes. §3.2 provides additional evidence against an aspectual classification of
the suffixes.
3.1.

-ahta/-aise are [+q] Diminutive Modifiers of Events

We argue that -aise/-ahta are [+quantized] diminutive modifiers of events (cf.
Armoskaite and Sherkina 2008, Progovac 2005, Wiltschko 2006) and that their
aspectual effects are epiphenomenal. The diminutive content of the suffixes is
sensitive to lexical aspect, here assumed to be temporal structure internal to an
event (Guéron and Lecarme 2004, Rothstein 2004).
Relying mainly on Rothstein (2004:12), we assume four aspectual classes.
States are events without either stages or endpoints (e.g. sleep, love). Activities
are events with stages, but lack endpoints (e.g. jump, run). Accomplishments
have both stages and endpoints (e.g. read a book, build a house). Finally,
achievements have only endpoints (e.g. arrive, win).
In what follows, we show that the Finnish suffixes interact differently with
the four aspectual classes, and that it is this interaction that gives rise to a
restricted range of interpretations. Recall that the lack of consistency in
distribution and motivation for interpretation make the traditional aspectual
classification of the suffixes problematic, as discussed in §1. The advantage of
our approach is that it allows us to predict both the distribution and the range of
interpretations. What the suffixes do is cut out a bounded portion of an event and,
thus, make it quantized, in the sense of Krifka (1998). However, the way the
quantization plays out in each aspectual class is subtly distinct.
When the diminutive [+q] applies to a stative predicate, the quantized event
is interpreted as a change of state, necessarily sudden, as (9) shows schematically.
The example in (10) illustrates the schematic model.
(9)

Schematic model I
Base V
…-------…
‘state’



V-ahta/-aise
]../..[
‘suddenly get into state’

5
(10)

a.

Lapsi nukku -i.
child sleep -PAST.3SG
‘The child was sleeping’

b.

Lapsi nuk -aht -i.
child sleep -ahta -PAST.3SG
‘The child fell asleep’

When the diminutive [+q] applies to an activity verb with discrete stages, the
quantized event is interpreted as one point event of the activity, necessarily
sudden and brief. The example in (12) illustrates the second schematic model.
(11)

Schematic model II
Base V
V-ahta/-aise
...--/--/--/--/--…

]../..[
‘discrete stages activity’
‘one brief & sudden instance of an activity’

(12)

a.

Lapsi hyppäs -i.
child jump -PAST.3SG
‘The child jumped’
Note: either once or many times

b.

Lapsi hyp -äht -i.
child jump -ahta -PAST.3SG
‘The child jumped’
Note: only once

When the diminutive [+q] applies to an accomplishment verb without
discrete stages, the quantized event is interpreted as necessarily brief (and
insignificant). The example in (14) illustrates the third schematic model.
(13)

Schematic model III
Base verb
V-ahta/-aise
…------->…

]../..[
‘non-discrete stages activity’ ‘a brief, non-chalant instance of the activity’

(14)

a.

Ignas luk -i
kirjan.
Ignas read -PAST.3SG book.ACC
‘Ignas read a/the book’

6
b.

Ignas luk -ais -i
kirjan.
Ignas read -aise -PAST.3SG book.ACC
‘Ignas read a/the book hastily, without paying much attention’

The diminutive [+q] does not apply to achievements, because achievements
are inherently quantized and minimal events.2 Examples in (16) illustrate the
ungrammaticality of model IV.
(15)

Schematic model IV
Base verb
]../..[
‘minimal point event’

(16)



Derived verb
n.a.

a.

*huom -ais -i
notice -aise -PAST.3SG

b.

*saav -aht -i
arrive -ahta -PAST.3SG

c.

*taju
-ais -i
realize -aise -PAST.3SG

d.

*tunt
-ais -i
recognize -aise -PAST.3SG

The table bellow summarizes the interpretations that arise based on the
interaction of the [+quantized] diminutive -ahta/-aise and the aspectual classes.
(17)
Aspectual Class
States
Activities
Accomplishments

[dim +q]
yes
yes
yes

Achievements

no

2

Interpretation
sudden change of state
event happened once, was brief & sudden
a brief event happened once suddenly &
briefly3
n.a.

Akin to how one can only have either ‘some’ or ‘a/the’ on a noun phrase, but not both.
(i)
Mary saw some man.
(ii)
Mary saw a man.
(iii)
*Mary saw some a man.
3
It also has the interpretation of being insignificant. We do not address this evaluative
interpretation here, but assume that it can be derived based on coercion of discrete stages
onto something that does not lend itself easily to discreteness of stages. Note that this is
the only subset that gets the pejorative interpretation without any contextual input.

7
Aspectual classification is organized based on the internal structure of
events. Once we take into account the internal structure of a given event, we can
predict the effect of the [+quantized] diminutive modifier -ahta/-aise. The
suffixes have aspectual effects in that the affixed predicates become quantized
events, but we argue that the aspectual effects are secondary and derivable form
the [+q] diminutive content.
Furthermore, if we treat the suffixes as [+quantized] diminutive
modifiers, their optionality and reference to seize/length of the event fall out from
their lexical specification. If the suffixes are treated as aspectual morphemes, the
optionality and reference to seize/length have to be stipulated, since these
qualities are not inherently aspectual, i.e. do not pertain to the internal structure of
an event.
3.2

Evidence Against Aspectual Label

In this section, we provide more evidence against the traditional aspectual
classification of -ahta/-aise.
Perhaps the strongest argument against the use of the aspectual label for
-ahta/-aise is the lack of consistency in the distribution of the suffixes within
aspectual classes. If -ahta/-aise were dedicated aspectual morphemes, they should
apply to any predicate within a given aspectual class (akin to how dedicated tense
morphemes apply to all verbs, dedicated determiners take all nouns, etc.). As
examples below indicate, this is not the case.
In (18) we see that some stative verbs are selected by either –ahta or -aise,
and others are ungrammatical with either.
(18)

States
a.

Lapsi nuk -aht -i.
child sleep -ahta -PAST.3SG
‘The child fell asleep’

b.

Ystävä viiv -äht-i
friend stay -ahta - PAST.3SG
‘The friend stopped by’

c.

*el -äht -i
live -ahta -PAST.3SG

/ *el -äis -i
live -aise -PAST.3SG

d. *tied -äht -i
know -ahta -PAST.3SG

/ *tiet -äis -i
/ know -aise -PAST.3SG

(19) illustrates that each of -ahta or -aise select for some activity verbs,
while others remain ungrammatical with either.

8
(19)

Activities
a.

Lapsi hyp -äht -i.
child jump -ahta -PAST.3SG
‘The child jumped once’

b.

Lapsi potk -ais -i
palloa.
child kick -aise -PAST.3SG ball. ACC
‘The child gave the ball a kick’

c.

*kiiv -äht -i
climb -ahta -PAST.3SG

/ * kiip -äis -i
climb -aise -PAST.3SG

d. *työnn –äht -i
push -ahta -PAST.3sg

/ *työnt -äis -i
push -aise - PAST.3sg

The data in (18) and (19) is unexpected and difficult to explain if the
morphemes are considered aspectual. Under our analysis, treating the suffixes as
[+ q] diminutive modifiers, this problem does not arise. Since modification is an
optional operation, application of the modifiers may be sensitive to particular
lexical content of the entities it aims to modify.
Furthermore, lexicalizations of the diminutive uses would be unexpected
of dedicated aspectual morphemes, but is allowed if -ahta/-aise are treated as
diminutive modifiers. Below we exemplify the use of -ahta/-aise in derivation of
new lexical entries.
(20)

4.

a.

häipy
-i
disappear- PAST.3SG
‘disappeared’

häiv
-äht -i
disappear -ahta -PAST.3SG
‘appeared fleetingly’

b.

puhu -i
speak -PAST.3SG
‘spoke’

puh -aht -i
speak -ahta -PAST.3SG
‘made a huffing sound’

Further Questions

In sections §2 and §3 we have made the case for the re-classification of -ahta/
-aise into [+q] diminutive modifiers and have shown that their aspectual effects
are secondary. However, pending questions still remain. In this section we
discuss problems that we have resolved only partially: sensitivity for argument
structure and the level of merge.

9
4.1

-ahta/-aise Sensitivity to Argument Structure

As a working hypothesis, we posit that -ahta/-aise may be sensitive to the
argument structure of the predicates to which they attach. Largely, this appears to
be true.
On one hand, -ahta selects for unaccusative predicates (ones underlyingly
without subject), as in (21) and does not allow transitive predicates, as in (22).
(21)

(22)

a.

kiehu -u
boil -PRES.3SG
‘(it) boils’

b.

kieh -ahta -a
boil -ahta -PRES.3SG
‘(it) comes to a boil’

a.

sito -o
solmun
tie - PRES.3SG knot.ACC
‘(s/he) ties a knot ’

b.

*sid -ahta -a
solmun
tie -ahta -PRES.3SG knot.ACC

-aise, on the other hand, selects for transitive predicates (those with subject
and object), as in (23) and does not permit unaccusative ones, as in (24).
(23)

(24)

a.

sito -o
solmun
tie -PRES.3SG knot.ACC
‘(s/he) ties a knot ’

b.

sit -aise -e
solmun
tie -aise -PRES.3SG knot.ACC
‘(s/he) quickly ties a knot’

a.

kiehu -u
boil -PRES.3SG
‘(it) boils’

b.

*kieh -aise -e
boil -aise -PRES.3SG

However, both -ahta and -aise select for subsets of unergative predicates

10
(ones underlyingly without subject), as shown in (25).
(25)

a.

naura -a
laugh -PRES.3SG
‘(s/he) laughs’

b.

naur -ahta -a
laugh -ahta -PRES.3SG
‘(s/he) gives a brief laugh’

c.

*naur -aise -e
laugh -aise - PRES.3SG

d.

pese -e
wash -PRES.3SG
‘(s/he) washes’

e.

pes -aise -e
wash -aise - PRES.3SG
‘(s/he) washes quickly’

f.

*pes -ahta -a
wash -ahta -PRES.3SG

We have yet to establish what factors influence the selection of certain
unergatives by -ahta, and what factors the selection of certain unergatives by
-aise. Our findings so far are summarized in the table below.
(26)
Type of Predicate
unaccusative +object, - subject
transitive +object, +subject
unergative (+object), + subject
4.2

-ahta
yes
no
yes

-aise
no
yes
yes

-ahta/-aise Level of Merge

We have classified -ahta/-aise as [+q] diminutive modifiers, but we have not
addressed their syntactic position. In this section we discuss evidence of merge
level of the two prefixes.
In Finnish, roots cannot stand on their own. As the contrast between (27) and
(28) indicates, formation of a grammatical verbal stem requires a vowel (bolded)
between the root and the inflection (Kiparsky 2003:146).

11
(27)

*kein, *vet
swing, pull

…

(28)

a.

kein -u
-u
swing -vowel -PRES.3SG
‘s/he swings’

b.

vet -ä
-ä
pull -vowel -PRES.3SG
‘s/he pulls’

We propose that the said vowel is an indication of a stem boundary.4 The
exact distribution of the vowel remains to be accounted for, but we tentatively use
the vowel as a test for affix merge level. With the exception of -ahta/-aise, the
vowel intervenes between all verbal suffixes and the root.
(29)

heil -u
-n
swing -vowel -PRES.1SG
‘I swing/sway/rock.’

(30)

heil
-u
-ile
-n
swing -vowel -FREQUENTATIVE -PRES.1SG
‘I keep swinging/swaying/rocking.’

(31)

heil
-u
-ta
-n
swing -vowel -CAUSE -PRES.1SG
‘I make (something) swing/sway/rock.’
Crucially, the vowel never intervenes between the root and -ahta/-aise.

(32)

4

a.

heil -ahda -n
swing -ahta -PRES.1SG
‘I swing/sway/rock once’

b.

pes -aise -n
wash -aise -PRES.1SG
‘I wash hastily’

c.

kein -ahta -a
swing -ahta -PRES.3SG
‘s/he swings once’

d.

vet -äise -e
pull -aise -PRES.3SG
‘s/he pulls once’

Kiparsky (2003) analyses stem and word level morpho-phonology, but does not discuss
root level morpho-phonology. Brattico (2005) discusses Finnish roots, but does not discuss
any phonological reflection of the root boundary.

12
The vowel does intervene between –ahta/-aise and other affixes.
(33)

heil
-ahd -u
-ta
-n
swing -ahta -vowel -CAUSE - PRES.1SG
‘I make swing/sway/rock once’

(34)

pes
-ais -u
-ta-n
wash -aise -vowel -CAUSE -PRES.1SG
‘I make (someone) wash (something) quickly’

(35)

heil
- ahd - u
- tt
-ele -n
swing -ahta -vowel -CAUSE -FREQ -PRES.1SG
‘I keep making (something) swing/sway/rock repeatedly’

Given the position of the vowel relative to -ahta/-aise, we tentatively
conclude that the two suffixes may be root level modifiers. In order to provide
more solid evidence, we first need to account for the distribution of the stem
vowel and find independent evidence for its role as a stem boundary indicator,
which is beyond the scope of this paper.
5.

Conclusion

In this paper, we have looked at two Finnish suffixes, -ahta/-aise. We have
questioned the traditional aspectual classification of these suffixes, based on the
lack of consistency in distribution and interpretation. We have argued that the
two suffixes are [+q] diminutive modifiers, since this allows us to: (i) account for
their role in the modification of event size/length; and (ii) predict their
distributive and interpretative peculiarities with regard to aspectual classes. We
have further shown that once the aspectual label is dropped, other lexical uses of
the suffixes are no longer surprising and fall out from the modifier status.
Although we have made progress in exploring how -ahta/-aise are
sensitive to the argument structure of the unaccusative and transitive predicates
they attach to, we have yet to provide a definitive analysis about their puzzling
interaction with unergatives. We have also made headway in establishing the
merge level of the suffixes and have good reason to speculate that -ahta/-aise are
root level modifiers. However, in order to complete the argument for root level
modification, we first need to gain better understanding of the structure of
Finnish stem.

13
References
Armoskaite, S., and M. Sherkina-Lieber. 2008. Event Number Suffixes in Russian and
Lithuanian. In Proceedings of the Annual Workshop on Formal Approaches to
Slavic Linguistics. Michigan Slavic Publications, 1-16.
Brattico, Pauli. A Category-Free Model of Finnish Derivational Morphology. 2005. SKY
Journal of Linguistics 18, 7-45.
Guéron , J., and J. Lecarme. (eds.). 2004. The Syntax of Time. Cambridge: MIT Press.
Hakulinen, Lauri. 1941. Suomen kielen rakenne ja kehitys. Helsinki: Otava.
Hakulinen, Auli, Maria Vilkuna, Riitta Korhonen, Vesa Koiviston, Tarja-Riitta Heinonen
and Irja Alho. 2004. Iso suomen kielioppi [ISK; The Great Finnish Grammar].
Helsinki: SKS.
Kangasma-Minn, Elina. 1984. Tense, aspect and Aktionsarten in Finno-Ugrian. In Aspect
Bound, eds. G. de Groot and Hannu Tommola. Dordrecht: Foris, 77-93.
Karlsson, Fred. 1983. Suomen kielen äänne- ja muotorakenne. Juva: WSOY.
Kiparsky, Paul. 2003. Finnish Noun Inflection. In Generative Approaches to Finnic and
Saami Linguistics, eds, D. Nelson and S. Manninen. CSLI Publications, 109-161.
Krifka, Manfred. 1998. The origins of telicity. In Events and Grammar, ed. S. Rothstein.
London: Kluwer, 197-235.
Progovac, L. 2005. A Syntax of Serbian: Clausal Architecture. Indiana University: Slavica
Publishers.
Rothstein, S. 2004. Structuring Events. Oxford: Blackwell.
Wiltschko, Martina. 2006. Why should diminutives count? In Organizing Grammar.
Linguistic Studies in Honor of Henk van Riemsdijk, eds. H. Broekhuis, N.
Corverm R. Huijbregtsm U. Kleinhenz and J. Koster. Berlin: Walter de
Gruyter, 669-679.