Tuesday, December 14, 2010


© 2010 by Taksin Nuoret

[en français]


Ever since December 2001, when the results of the first PISA survey were made public, the Finnish educational system has received a lot of international attention. Foreign delegations are flocking to Finland, in the hope of discovering Finland's secrets.

The explanation widely accepted is that the Finnish educational system is better. For example, the following aspects have been pointed out:
  • Schools routinely provide tutoring for weak students.
  • Each school has a social worker ("koulukuraattori").
  • Substitute teachers are often provided when the teacher is ill.
  • Teachers are seldom on strike.
  • The methods used for teaching mother tongue are solid. Finnish first graders learn to read first by learning letters, then syllables, then words, then sentences. For example, throughout grade 1 (and most of grade 2), words are often printed with syllables separated by hyphens [1]. Adventurous approaches (such as starting with words or sentences as wholes) are not used.
  • Schools have more autonomy than in many countries. For example, schools can dismiss teachers if they are not satisfied with their work.
(As a parent of children who attend Finnish schools, I can confirm that the above points are true, although I am not going here to argue about their relevance.)

Explanations not related to the educational system have also been proposed, including:
  • The Finnish society is homogeneous. The number of foreigners is lower than in most OECD countries (2,9% at the end of 2009 [2]), which makes the teachers' job easier.
  • Finnish spelling is regular, thus easing Finnish schoolchildren's task.
  • Foreign TV programs are subtitled, instead of dubbed as in many OECD countries, thus easing acquisition of foreign languages.

Although explanations not related to the educational system are mentioned now and then, the ones inherent to the Finnish educational system have been predominantly put forward by the media, both inside and outside Finland. [3a] and [3b] are typical examples in this respect.

Many people in Finland, including myself, have been willing to accept the comfortable explanation that the Finnish PISA success was due mainly to its educational system.

But then, at the end of 2007 the results of PISA 2006 were made public...

The case of Estonia

In 2006, PISA was conducted for the third time. In 2000 and 2003, mostly OECD countries took part to PISA, but in 2006 the group of so-called "partner country economies" taking part was greatly extended, and included countries as varied as Brazil, Bulgaria, Croatia, Estonia, Israel, Latvia, Qatar, Romania, Thailand and Uruguay. Altogether 57 countries took part in 2006 [4].

As in the 2000 and 2003 surveys, once more Finland got impressive results, well ahead of the pack - see Table 2 of [4].

A striking aspect of the 2006 results was how well Estonia performed: 5th place worldwide (after Finland, Hong Kong, Canada and Taiwan). That is, second country in Europe after Finland - see Table 2 of [4]. In Table 1 of [4], Estonia is even listed second country worldwide (after Finland).

Estonia is a small country (1,34 million inhabitants), with turbulent recent history (it restored its independence only in 1991). Its GDP is modest compared to that of many OECD countries.

Now, isn't the fact that such a country got better PISA results than all European OECD countries (except Finland) at least as remarkable as Finland's success? Could there be a common factor behind the success of Finland and that of Estonia?

There is a common factor: language.

Finnish and Estonian are not Indo-European, but Finno-Ugric languages. Hungarian also belongs to the family of Finno-Ugric languages. Finnish and Estonian belong to the Finnic branch of the Finno-Ugric family, whereas Hungarian is the main representative of the Ugric branch.

Finnish and Estonian are so closely related that it is possible for Finns who have never studied Estonian to understand Estonian to some extent, and vice-versa. At least many individual words can be recognized, even though overall understanding is often challenging.

So, do Finland and Estonia both get top PISA results by some remarkable coincidence, or is language playing a key role?

The case of Swedish-speaking Finns

There is actually a way to find out whether Finnish language plays a key role in Finland's success.

Finland has two official languages: Finnish and Swedish [5]. At the end of 2009, the Swedish-speaking minority represented 5,43% of the Finnish population, the Finnish-speaking majority 90,67% [6].

The Swedish-speaking population is concentrated on the western and southern coasts of the country. The archipelago of Åland, located between Finland and Sweden, is almost exclusively Swedish-speaking.

For historical reasons, the proportion of Swedish-speaking Finns in higher classes of the Finnish society has been high. This was true especially until the beginning of the 20th century, but even today noticeable differences remain. For example:
  • At the beginning of 1997, 9,1% of Finnish-speaking Finns, but 14,1% of Swedish-speaking Finns owned stocks [7].
  • At the beginning of 1997, among those who owned stocks, the mean investment wealth was 69700 FIM [8] amongst Finnish-speaking Finns, but 221100 FIM among Swedish-speaking Finns [7].

The Swedish-speaking population, in addition to being wealthier on average than the Finnish-speaking population, is also said to enjoy a richer social life, to have better self-esteem, to be more tolerant, to have a much higher life expectancy, etc. [9,10]

In Finland, too, socio-economic background of students greatly influences school results (although less than in most other OECD countries). Excerpt from [11] (p. 35):

    "Students whose parents had the highest status jobs significantly outperformed those with lower socio-economic backgrounds. This was especially the case in, for instance, Hungary, Belgium, Turkey and Germany. The difference was considerable in Finland as well, yet remained clearly below the OECD average (Figure 13)."

Therefore, one could expect that Swedish-speaking Finns would get considerably better PISA results than Finnish-speaking Finns.

There is indeed a difference between PISA results of Swedish-speaking and Finnish-speaking students, but opposite to the expected one. Excerpt from [11] (p. 17):

    "In PISA 2003 Finnish-speaking students clearly outperformed their Swedish-speaking peers in scientific literacy, with an average difference of 26 points. However, also the Swedish-speaking minority was doing very well, since their results were on a par with those of the Netherlands."

(Let's take the opportunity provided by this quote to note that if Finland's population happened to be mostly Swedish-speaking, the Finnish educational system would not be the focus of international attention.)

Same country, same Ministry of Education, inferior socio-economic background on average, yet superior PISA results. The question is not whether the Finnish language is a key factor to Finland's PISA results, but why.

Side notes on Swedish-speaking students in PISA

20,8% of the Finnish students who took part in PISA 2003 were Swedish-speaking, that is, much more than the share of the Swedish-speaking population (which was 5,55% in 2003 [6]). Excerpt from [12]:

    "In Finland the PISA 2003 sample comprised 147 Finnish-language schools and 50 Swedish-language schools. The population was 6,235 students, of whom 5,796 (93%) answered the test questions. Of these 4,589 were Finnish-speakers and 1,207 Swedish-speakers."

This means that the PISA 2003 results of the Finnish-speaking students are actually even higher than those reported for the whole country (since the results reported for the whole country include results of both linguistic communities).

In PISA 2006, much less Swedish-speaking students took part: only 5,7%, which is close to the share of the Swedish-speaking population (5,49% in 2006 [6]). Translated excerpt from [13]:

    "In Finland the PISA 2006 sample comprised 144 Finnish-language schools and 11 Swedish-language schools. The population was 5,265 students, of whom 4,714 (90%) answered the test questions. Of these 4,413 were Finnish-speakers and 301 Swedish-speakers."

I was unable to find corresponding numbers for PISA 2000.

Why does Finnish give better PISA results?

As briefly mentioned in the introduction, Finnish spelling is regular. That is, the correspondence between letters and phonemes is tight: in general, one phoneme corresponds to one letter, and one letter to one phoneme. Thus, a Finn who hears an unknown word knows how to spell it, and a Finn who reads an unknown word knows how to pronounce it. There are some exceptions, though [14].

Estonian spelling is about as regular as Finnish spelling.

So, it seems clear that regular spelling is the decisive factor, doesn't it?

One must admit that, everything else being equal, regular spelling helps: schoolchildren don't have to spend much of their curriculum to learn how to read and write correctly.

But then, in PISA Spain and Italy are far behind such English-speaking countries as New Zealand, Australia, UK and Ireland [4]. And this despite the fact that Spanish and (especially) Italian spellings are much more regular than the intricate English spelling [15,16]. Thus, correlation between regular spelling and PISA results is weak.

If regular spelling explains only to a little extent the advantage given by Finnish, which other aspects of the language could be involved?

Finnish morphology

As many languages, Finnish can form new words from roots in two ways: composition and derivation. Examples in English:
  • [composition] 'rail' + 'way' = 'railway' (two lexemes combine to form a new lexeme)
  • [derivation] 'bitter' + '-ness' = 'bitterness' (a lexeme combines with an affix to form a new lexeme)

One remarkable aspect of Finnish is the richness of its derivational morphology. Compositional morphology, too, is very rich in Finnish, but so it is in many languages of countries who took part in PISA.

Here is an example of the richness of Finnish derivation:
  • 'kirja' - 'book'
  • 'kirjailija' - 'writer'
  • 'kirjailla' - 'to embroider'
  • 'kirjailu' - 'embroidery'
  • 'kirjaimellinen' - 'literal'
  • 'kirjaimellisesti' - 'literally'
  • 'kirjaimisto' - 'alphabet'
  • 'kirjain' - 'letter' (of the alphabet)
  • 'kirjallinen' - 'written; literary'
  • 'kirjallisuus' - 'literature'
  • 'kirjaltaja' - 'typographer'
  • 'kirjanen' - 'booklet'
  • 'kirjasin' - 'font' (in typography)
  • 'kirjasto' - 'library'
  • 'kirjata' - 'to write down, to make a note of'
  • 'kirje' - 'letter' (document)
  • 'kirjeellinen' - 'by letter' (adj.)
  • 'kirjelmä' - 'letter, note, message'
  • 'kirjelmöidä' - 'to complain (by means of writing)'
  • 'kirjoitella' - 'to write' (now and then)
  • 'kirjoittaa' - 'to write'
  • 'kirjoittaja' - 'writer' (person who performs a writing work)
  • 'kirjoittaminen' - 'writing' (action)
  • 'kirjoittautua' - 'to get enrolled, to log in'
  • 'kirjoittelu' - 'writing' (now and then)
  • 'kirjoitus' - 'writing' (result)
  • 'kirjoituttaa' - 'to have ... written' (factitive of 'kirjoittaa')
  • 'kirjuri' - 'scribe'
(And then, there is naturally a myriad of words obtained by composition that use one of the above lexemes.)

Sections 53, 54, 55, 56, 58, 59, 66, 67, 68 and 69 of [17] give impressive lists of Finnish suffixes.

Thus, the number of roots needed to reach comparable vocabulary is lower in Finnish than in, say, English, Spanish, French or Italian.

Another aspect of Finnish morphology worth noting is the fact that foreign affixes and stems (for example, Greek and Latin affixes and stems) are used less often in Finnish than in many languages of OECD countries. And when Greek or Latin affixes or roots are used in Finnish, the words thus obtained are often used in parallel to other Finnish words that belong to older layers of the Finnish lexicon.

Compare, for example:
  • [Finnish] 'kaatumatauti' ('kaatua' = 'to fall', 'tauti' = 'disease')
  • [English] 'epilepsy'
A Finnish speaker who hears 'kaatumatauti' for the first time understands the word. An English speaker who hears 'epilepsy' for the first time probably doesn't (if he or she does not know Greek). (Note that 'epilepsia' is also used in modern Finnish.)

Other example (with Latin this time):
  • [Finnish] 'kyynelpussi' ('kyynel' = 'tear', 'pussi' = 'sac')
  • [English] 'lachrymal sac'

The relatively rare use of stems and affixes of foreign origin makes Finnish morphology more transparent than the morphology of most languages of OECD countries.

The authors of [18] conclude their abstract with:

    "At any rate, the results suggest that Finnish elementary school children benefit significantly from utilizing morphology in determining word meanings."


What about science?

Finland achieved top PISA results not only in 2000 (when the focus of the study was on reading literacy), but also in 2003 (focus on mathematical literacy) and in 2006 (focus on science literacy).

It seems clear that the advantage given by the Finnish language explains not only Finland's PISA 2000 results, but also its 2003 and 2006 results, given that mathematical and science tasks require good reading comprehension in the first place.

The same remarks made above about transparency of Finnish morphology naturally apply just as well to mathematical and scientific terms.

Compare, for example:
  • [Finnish] 'viisikulmio' ('viisi' = 'five', 'kulma' = 'angle'; 'viisi' and 'kulma' are Finnish words)
  • [English] 'pentagon' ('pente' = 'five', 'gônia' = 'angle'... in Greek)
Transparent morphology definitely helps also in maths and in science.

Unfortunately, I have been able to find sample PISA questions only in English [19]. It would be interesting to compare the English version and the Finnish translation of the questions, especially in mathematics and science.

Conclusion and further work

Finland's PISA success probably results from the complex combination of several factors.

Comparisons between the results of Finnish-speaking and Swedish-speaking students in Finland show that language plays a role that is more important than is suggested by most commentators. This observation is reinforced by Estonia's remarkable PISA results.

I have mentioned possible explanations as to why the Finnish language gives an edge: transparent spelling, rich derivational morphology, and transparent morphology.

I have only seen transparent spelling mentioned by commentators (and usually only as a secondary explanation to Finland's success). I have tried to show why I believe that rich derivational morphology and transparent morphology play a more important role than transparent spelling.

It would be interesting to see more studies on the impact of (Finnish) morphology on learning.


[1] See, for example, the book Salainen aapinen (WSOY, 2000), used in grade 1, where one can find dialogs such as (p. 66):
- Si-nul-la on ki-va ve-li.
- Sen ni-mi on Vil-le.
- Voi-ko Vil-len ot-taa sy-liin?

[2] Statistics Finland, http://www.stat.fi/tup/suoluk/suoluk_vaesto_en.html#foreigners

[3a] Why do Finland's schools get the best results?, http://news.bbc.co.uk/2/hi/programmes/world_news_america/8601207.stm

[3b] Une éducation finlandaise, http://www.youtube.com/watch?v=JBAPf0DFp1s (in French)

[4] The Programme for International Student Assessment (PISA), http://www.pisa.oecd.org/dataoecd/15/13/39725224.pdf

[5] In addition, Saami (spoken in Lapland) is recognised as a regional language. There are three variants of Saami spoken in Finland: Northern Saami (about 3000 speakers in Finland), Skolt Sami (400 speakers in Finland, very few speakers outside Finland), Inari Saami (300 speakers, all in Finland). More details are available at http://www.ethnologue.com.

[6] Statistics Finland, http://www.stat.fi/tup/suoluk/suoluk_vaesto_en.html#structure

[7] Shareownership in Finland, Matti Ilmanen and Matti Keloharju, http://lta.hse.fi/1999/3/lta_1999_03_a3.pdf

[8] FIM = Finnish mark. Finland took the euro into everyday use on 1st January 2002. 1 euro = 5,94573 Finnish marks.

[9] Suomenruotsalaisten pidemmän iän salaisuus ei ole raha vaan kirjastokortti, http://ylex.yle.fi/radio/ohjelmat/ylex-tanaan/mielipidevanki/suomenruotsalaisten-pidemman-ian-salaisuus-ei-ole-raha-vaa (in Finnish)

[10] Suomenruotsalaiset elävät muita pidempään - miksi?, http://www.finland.se/public/default.aspx?contentid=114462&nodeid=36125&contentlan=1&culture=fi-FI (in Finnish)

[11] The Finnish success in PISA - and some reasons behind it - PISA 2003, Institute for Educational Research, University of Jyväskylä, http://ktl.jyu.fi/ktl/julkaisut/luettelo/vuosi_2007/d084, http://ktl.jyu.fi/img/portal/8317/PISA_2003_print.pdf

[12] OECD PISA 2003: Young Finns among the World Top in Learning Outcome, Ministry of Education and Culture, http://www.minedu.fi/OPM/Tiedotteet/2004/12/oecdn_pisa_2003_-tutkimus_suomalaisnuorten_osaaminen_maailman_?lang=&extra_locale=en

[13] PISA 2006, Ministry of Education and Culture, http://www.minedu.fi/OPM/Koulutus/artikkelit/pisa-tutkimus/pisa2006 (in Finnish)

[14] Let's mention these ones:
  • Assimilation phenomena are not rendered in spelling. For example, 'olenpa' ('I am indeed') is pronounced 'olempa'.
  • There is no letter for the velar nasal. For example, 'kenkä' ('shoe') is pronounced 'keŋkä' (not 'kenkä'), and 'kengän' ('shoe', genitive singular) is pronounced 'keŋŋän' (not 'kengän').
  • Final doubling ('loppukahdennus'). For example, 'tervetuloa' ('welcome') is pronounced 'tervettuloa' in standard Finnish.
  • Isolated exceptions. For example, 'sydämen' ('heart', genitive singular) is pronounced 'sydämmen'.
[15] The Transparency of Spanish Orthography, Ian Mackenzie, http://www.spellingsociety.org/journals/j15/spanish.php

[16] Italian spelling, and how it treats English loanwords, Christopher Upward and Virginia Pulcini, http://www.spellingsociety.org/journals/j20/italian.php

[17] Suomen kielen rakenne ja kehitys. Lauri Hakulinen, Otava, 1979 (in Finnish)

[18] The role of derivational morphology in vocabulary acquisition: Get by with a little help from my morpheme friends, Raymond Bertram, Matti Laine and Minna Maria Virkkala, http://onlinelibrary.wiley.com/doi/10.1111/1467-9450.00201/abstract

[19] Take the Test - Sample Questions from OECD's PISA Assessments, http://www.oecd.org/dataoecd/47/23/41943106.pdf

© 2010 by Taksin Nuoret