Wednesday, October 30, 2013

© 2013 by Taksin Nuoret

[en français]

Table of Contents

Changes to the previous version

In addition to minor changes and clarifications to the previous version, I have tried here to provide more information on previous studies on the impact of language, Finnish in particular, on learning processes.

Introduction

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. Finland is also trying to take advantage of its PISA success by exporting its knowledge in education [1]; this strategy is supported by talks given in international events by representatives of the Finnish Ministry of Education and Culture [2].

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 [3]. 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.
  • The profession of teacher is better recognized than in many countries.
  • Transition from low to high grades of the Finnish curriculum is smoother than in many countries.
  • Finnish students have a free canteen at their disposal.
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 (3.6% at the end of 2012 [4]), which makes the teachers' job easier.
  • Finnish spelling is regular, thus easing Finnish students' task.
  • Foreign TV programs are subtitled, instead of dubbed as in many OECD countries, thus easing acquisition of foreign languages.
(Note that I am not claiming here that these aspects explain or do not explain Finland's success, I am only reporting that they have been used by some as possible explanations.)

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. [5a] and [5b] are typical examples. And many people have been willing to accept the explanation that Finland's PISA success is 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 [6].

As in the 2000 and 2003 surveys, once more Finland got impressive results, well ahead of the pack - see Table 2 of [6].

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 [6]. In Table 1 of [6], Estonia is even listed second country worldwide (after Finland).

Estonia is a small country (1.29 million inhabitants in 2011), with turbulent recent history (it restored its independence only in 1991). Its GDP per capita 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 an important role?

The case of Swedish-speaking Finns

There is actually a way to further investigate whether the Finnish language plays an important role in Finland's success.

Finland has two official languages: Finnish and Swedish [7]. At the end of 2012, the Swedish-speaking minority represented 5.36% of the Finnish population, the Finnish-speaking majority 89.68% [8].

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. Let's quote [9]:

    "3.4. Investment wealth and mother tongue

    Table 7 investigates how mother tongue is related to investment wealth. The Swedish-speaking minority (5.8% of the Finnish population) is much wealthier than the Finnish-speaking majority (92.9% of population): the average investment wealth of Finnish-speaking Finns owning stocks, 69,700 FIM, is less than one-third of the investment wealth of Swedish-speaking Finns owning stocks, 221,100 FIM. The ratio of investor-inhabitants to all inhabitants is also greater for Swedish-speaking Finns (14.1%) than for Finnish-speaking Finns (9.1%). Therefore, the value of the stock portfolio of an average Swedish-speaking Finn is more than four times as large as that of an average Finnish-speaking Finn."

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 [11], etc.

There is by the way a spiteful Swedish saying against the Swedish-speaking minority, still used sometimes by some Finnish-speakers in Finland: "svenska talande bättre folk" ("the better Swedish-speaking people").

In Finland, too, socioeconomic background of students greatly influences school results (although less than in most other OECD countries). Excerpt from [12] (p. 35):

    "Students whose parents had the highest status jobs significantly outperformed those with lower socioeconomic 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, who are tested in Swedish, would get significantly better PISA results than Finnish-speaking Finns.

There is indeed a difference between PISA results of Swedish-speaking and Finnish-speaking Finnish students, but opposite to the expected one. Excerpt from [12] (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."

The PISA results of the Netherlands are indeed not bad at all, but 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 probably not be the focus of international attention — at least not more so than the Dutch educational system.

Same country, same Ministry of Education, inferior socioeconomic background on average, yet superior PISA results. Why does Finnish as a native language give better PISA results?

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 [8]). Excerpt from [13]:

    "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 [8]). Translated excerpt from [14]:

    "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.

Finnish orthography

Finnish spelling is very 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 only a few exceptions [15].

Estonian spelling is about as regular as Finnish spelling.

It is clear that, everything else being equal, regular spelling helps: schoolchildren don't have to spend as much of their curriculum to learn how to read and write.

There are, however, other aspects of the language which help.

Finnish morphology

As many languages, Finnish can form 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, for example in German, Dutch or Swedish.

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 [16] give impressive lists of Finnish suffixes.

Thus, the number of roots needed to reach a comparable vocabulary is lower in Finnish than in many other languages.

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] 'saaristo' ('saari' = 'island', '-sto' = collective suffix, 'group of')
and:
  • [English] 'archipelago'
A Finnish speaker who hears 'saaristo' for the first time understands the word. An English speaker who hears 'archipelago' for the first time probably doesn't (if he or she does not know Greek).

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

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).

Given that mathematical and scientific tasks require good reading comprehension in the first place, it seems clear that any advantage given by the Finnish language can explain not only Finland's PISA 2000 results, but also its 2003 and 2006 results.

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)
and:
  • [English] 'pentagon' ('pente' = 'five', 'gônia' = 'angle'... in Greek)
Transparent morphology definitely helps also in mathematics and science. And that help comes in addition to the fact that Finnish-speaking schoolchildren, not having to spend as much of their curriculum to learn how to read and write, have more time to learn other topics, for example mathematics and science.

Unfortunately, I have been able to find sample PISA questions only in English [17]. It would be interesting to compare the English version and the Finnish translation of the questions, especially in mathematics and science.

Previous studies

In an article published in 2000 [18], the authors report their findings in experiments they conducted with Finnish-speaking third and sixth graders. They made several interesting discoveries regarding how morphological factors influence vocabulary acquisition.

Rather than attempting to summarize the article, I prefer to cite it verbatim:
  • [p. 288] "Finnish is a Finno-Ugric, agglutinative language with a very rich morphology."

  • [p. 288] "Moreover, derivation and compounding is very productive in Finnish, leading to huge morphological families."

  • [p. 288] "To summarize, the bulk of Finnish words in running text is polymorphemic (for nouns, 97.4%) and of very low frequency."

  • [p. 292] "More interestingly, the morphological make-up of words clearly affects our subjects' performance in giving word definitions. First of all, overall performance is poorest for the category for which no fall-back on morphology is possible, that is, the monomorphemic words, and best for derived words in high-productive suffixes."

  • [p. 292] "However, for derived words, sub-lexical units do exist in the form of rather high-frequent morphemes, and calculating the meaning on the basis of these units provides a rather successful back-up option, be it more successful for high- than for low-productive derivations."

  • [p. 294] "This root frequency effect would be yet another line of evidence that at the low-frequency range children effectively compute the syntax/semantics of complex words on the basis of the constituent morphemes."

  • [p. 294] "In that respect it does not come as a surprise that young Finnish children are more sensitive to word-internal morphological structure than their Indo-European age-mates."

  • [p. 294] "Thus frequency, suffix productivity and language all seem to affect the role that morphological structure plays in vocabulary acquisition."
The authors conclude their article as thus:
  • [p. 294] "The more general notion of this study is that children benefit greatly from utilizing morphology in determining word meanings. This might be particularly handy while they are engaged in listening to speech or in reading texts with a high number of infrequently used words. As shown in this study, it is especially in the low-frequency range that they will get by with a little help from their morphemic friends."
The interested reader is invited to read the article in its entirety. Unfortunately, only the abstract is available online.

Other studies have been made on the impact of language (in general) on several aspects of learning, and of Finnish in particular. Let us briefly mention two of them, where the authors emphasize the advantage given by a transparent orthography:
  • [19, p. 23] "Among thirteen European orthographies Finnish has the most shallow orthography and simplest syllabic structure, whereas the most complex orthography is found in English. This likely explains the high probability that Finnish children will achieve accurate and relatively fluent word reading skill before the end of the first school year."

  • [conclusion of 20, p. 168-169] "According to this interpretation, the slow rate of foundation literacy acquisition by the Scottish sample can be seen as an inevitable consequence of the complexity of the orthography and phonology of English. It may be that the rate of learning can be influenced at the margins, being further delayed by socioeconomic disadvantage (Duncan & Seymour, 2000), possibly accelerated by modifications to the teaching of phonics, and perhaps sensitive to the child's cognitive maturity when the teaching of reading is introduced. However, even where these educational aspects are all optimal, there will always be a cost associated with the implementation of a dual foundation process, and this will create irreducible differences in rates of progress between learning to read in English or other deep orthographies and learning to read in a shallow orthography."
Conclusion and further work

Finnish-speaking Finnish students get better PISA results than Swedish-speaking Finnish students, against all odds set by socioeconomic factors. This is a strong sign that language plays a role that is more important than suggested by most commentators.

Scientific papers have shown there to be an influence of language on various aspects of learning. The three papers cited above show that Finnish possesses several characteristics that help learning. These characteristics include: transparent orthography, simple syllabic structure, very rich (derivational) morphology.

Therefore, it would be as preposterous to claim that Finnish plays no role in Finland's success as it would be to claim that Finnish alone explains everything. The PISA results of any country are necessarily due to a complex interaction of several factors.

In order to better understand the nature and extent of the advantage given by Finnish, it would be interesting to carry out comparative studies between the Finnish-speaking and Swedish-speaking Finnish populations, for example in vocabulary acquisition by children.

References

[1] Future Learning Finland, http://www.futurelearningfinland.fi

[2] For example, go to http://www.youtube.com and search for "Pasi Sahlberg".

[3] 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?
(etc.)

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

[5a] Why do Finland's schools get the best results?, http://news.bbc.co.uk/2/hi/programmes/world_news_america/8601207.stm

[5b] Une éducation finlandaise, http://www.youtube.com/watch?v=JBAPf0DFp1s (in French)

[6] The Programme for International Student Assessment (PISA), http://www.oecd.org/pisa/pisaproducts/pisa2006/39725224.pdf

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

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

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

[10] FIM = Finnish mark. Finland took the euro into everyday use on 1st January 2002. 1 euro = 5.94573 Finnish marks.

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

[12] 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

[13] 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

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

[15] 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'.
[16] Suomen kielen rakenne ja kehitys. Lauri Hakulinen, Otava, 1979 (in Finnish)

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

[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, Scandinavian Journal of Psychology, 2000, http://onlinelibrary.wiley.com/doi/10.1111/1467-9450.00201/abstract

[19] Learning to Read: Reciprocal Processes and Individual Pathways, Marja-Kristiina Lerkkanen, PhD thesis, 2003, https://jyx.jyu.fi/dspace/bitstream/handle/123456789/13303/9513917827.pdf

[20] Foundation literacy acquisition in European orthographies, Philip H. K. Seymour, Mikko Aro and Jane M. Erskine, British Journal of Psychology, 2003, http://lyddansk.dk/sites/default/files/files/Foundation%20literacy%20acquisition%20in%20European%20orthographies.pdf

© 2013 by Taksin Nuoret