Text: Terhi Hautamäki
Illustrations: Kiia Beilinson
Retired university lecturer Heikki Lokki has shared his own speech with researchers for several decades.
In the early 1970s, he participated in the collection of colloquial Finnish spoken in Helsinki, representing a young speaker in his twenties. He participated again in the 1990s when he was in his forties and in the 2010s in his sixties. Each time that he sat down to be interviewed by research assistants who were university students of Finnish, he talked for about an hour while his speech was recorded.
Lokki no longer remembers the topics of the interviews; after all, it was quite a long time ago. Most likely he was encouraged to talk about his childhood memories, hobbies, work, living in Helsinki and perhaps his own observations about Finnish. These are the topics that the interviewers posed questions about to the interviewees. During the collection, the speakers were encouraged to talk about topics that were an everyday part of their lives and that they could talk about at length without struggling to find things to say.
Carried out for the first time in the 70s, interviewees were sought for the project based on demographic data. The linguistic corpus consists of interviews of more than two hundred people in total on three separate occasions, with some of the interviewees taking part twice or three times.
Lokki himself thinks his speech has changed as he has grown older. At the very least, he has gained new vocabulary both at work and in his free time. He has built a career as a lecturer on data processing and has been an avid bird-watcher.
“You don’t see the gradual changes yourself. That’s why this longitudinal study is interesting. My attitude towards the study has been positive: if I can be of some benefit to the research I can always find some time for it,” Lokki says.
Differences in spoken Finnish are diminishing
The speech corpus has provided fascinating information, for example, on how the differences in the speech of different population groups are diminishing. The first time that spoken Finnish speech was collected, the interviewees included well-spoken speakers of standard Finnish, while others spoke very colloquial Finnish. The differences between people from different backgrounds were much more obvious than they are today.
“Substantially codified standard Finnish is clearly a phenomenon that is about to disappear if it hasn’t already,” says University Lecturer, Docent Unni Leino from Tampere University who leads the Kone Foundation-funded Kippo project.
The project is called The Temporal Variation of Linguistic Populations and it uses the same corpus of spoken Finnishfor which Heikki Lokki has provided his speech.
The same corpus has been used to study variations in phonological and morphological structures, the slang of young people from different generations and the use of place names. The current project led by Leino involves interdisciplinary cooperation: the linguists involved seek speaker populations with the help of tools used by biologists and study temporal variations in linguistic populations. In this study, a speaker population refers to a group of speakers who resemble each other by linguistic features and differ from other speaker groups.
The speaker’s age and the decade in question account for the variation to some degree, but the significance of their place of residence and social class has declined. No social conclusions can be immediately drawn from this, but it is interesting to consider whether this reflects a move towards a more equal society on some level.
Social change has been evident already during the collection of the corpus and in the underlying hypotheses. In the 1970s, it made sense to look for representatives of two different social classes and to include highly educated people from the district of Töölö and people from a working-class background from the Kallio-Sörnäinen area. When collecting material in the 2010s, this division no longer made sense, as Kallio had become an expensive and trendy residential area for well-educated people. If the project continues at the same pace, the next material collection will take place in the early 2030s. Because people live and stay healthy longer, the previous collection already included a new age group: those over 80. It is possible that, in the future, the material will include people whose speech has been collected for up to 60 years.
Ordinary speakers may not consciously think about the social background of the people they talk to. But this does not mean that it doesn’t matter. Finnish speakers make unconscious deductions on where a person comes from based on the length of their vowels, for example. They may also make assumptions about a person’s education and occupational status based on their choice of words or colloquialisms.
Is it possible to study variations in an individual’s speech?
It’s not important to Unni Leino whether the interviewees have talked about their horse riding hobby or their work as a mathematics teacher or whether they have reminisced about the war. She is interested in phonological and morphological features and the computational study of these with the help of computer software.
In fact, Leino hasn’t even read through the interview transcripts. She only found out during the interview for this article that she has been studying the speech of her former teacher and colleague! You see, Leino started her career as a researcher in computer science at the University of Helsinki where Heikki Lokki worked at the time.
The original subject of Leino’s doctoral thesis was the processing and analysis of spatial data. Her material consisted of the National Land Survey of Finland’s spatial data register and the nomenclature of the entire base map.
As it happens, place names gradually became more fascinating to her than data processing. When her doctoral thesis was halfway finished, she found she had drifted off track and consequently applied for the right to further studies in Finnish, in which she had minored. In the end, she completed her doctoral thesis for the Department of Finnish.
Once, when she was visiting her former workplace at the Institute for the Languages of Finland, she bumped into a colleague who told her about an interesting project. It involved using the computational methods of biology to study language variation and differentiation. Leino ended up joining the project.
Cooperation between linguists and biologists may sound peculiar at first, but the collaboration has worked well in Finland for more than a decade. Many researchers from different universities and sectors have been involved in a family of research called BEDLAN, which is short for Biological Evolution and the Diversification of Languages. The Kippo project is also part of the BEDLAN project family.
The projects have focused on variation in Finno-Ugric languages and Finnish dialects. Some years ago, the researchers wondered whether the period covered could be shrunk even further and whether the same methods could be used to study material from an even shorter period, i.e. variation in the spoken language of individuals.
Language change brings up strong emotions in a lot of people
The Kippo project has focused on the study of certain variants in speech, i.e. the parallel forms of words and their variations. The research team is interested in change both during an individual person’s lifespan and at the community level.
To a layperson’s ears, such matters of form may sound insignificant. In an everyday conversation, it hardly matters what word form you use as long as the people listening understand you. But for Leino and other linguists, different forms of words provide fascinating information about how and why language changes with time. Strictly speaking, language change can only be seen with time, but by examining the variations occurring in language you can get a clue of the direction of the change.
Linguists are interested in studying language change because language is a key part of our culture. Language variation and change reflect social change and partly also influence it.
After closer examination, it is evident that changes in language at the level of form interest laypeople too. Take, for example, the Institute for the Languages of Finland, which is devoted to the study and language planning of Finnish and Swedish, the country’s official languages. Many Finns probably remember the strong emotional reactions the decisions made by the institute have caused over the years when common expressions that were previously considered incorrect were accepted among the recommended forms. One such decision in 2014 was made after a close voting result, and after its publication, the chairperson’s phone wouldn’t stop ringing, as irate citizens wanted to let her know what they thought about the decision. The media and debaters on social media made a big deal about it, and the newspaper Helsingin Sanomat, for example, announced that it would stick to the previous norm.
Biological species diverge, and so do languages
Evolutionary biology studies genetic variation within species. This variation may, in time, lead to several different species diverging from the original one. Similar variation occurs within a language which may, in time, result in branches of the language developing into different dialects and languages.
“It’s not a new idea that biological evolution and language divergence have something in common. Language family trees have been compiled since the 19th century,” Leino says.
In the past, researchers used to try to find a fundamental connection and similarities between the evolution of species and that of languages. According to Leino, they no longer try to make a direct analogy but take a pragmatic approach instead. Projects involving linguists and biologists are seeking ways to borrow useful computational tools from each other’s disciplines for analysing very different kinds of material. Different methods can provide new explanations on language variation and change or offer new ways of making earlier models more precise.
“We are looking into whether these similarities between biology and language are of the type that tools developed by biologists could be applied to linguistics. For this purpose, you need to know both the method and its scope of application, which requires a multidisciplinary research team.”
The Kippo project on spoken language involves postdoc researcher Jenni Santaharju from the University of Helsinki, whose doctoral thesis on evolutionary biology examines the divergence of ant populations. In 2014, she dove into very different kind of material after learning that the BEDLAN team of linguists needed an expert in population genetics calculation.
“Biologists have developed a large number of quantitative tools that are easy to use and that can be used to quickly analyse large volumes of material,” Santaharju explains.
In addition to Leino and Santaharju, the Kippo project includes Olli Kuparinen from Tampere University, who recently submitted his doctoral thesis for public examination, University Lecturer of Finnish Liisa Mustanoja and Professor of Statistics Jaakko Peltonen.
Santaharju finds the spoken Finnish of Helsinki fascinating because of its historical background: it has been influenced, for example, by Swedish and Russian. Unlike the Finnish spoken in many other towns and cities in Finland, Helsinki’s spoken Finnish is not based on a single Finnish dialect. Instead, it is a combination of many as a result of migration to the capital, especially dialects spoken in the Häme (Tavastia) and Southern Häme regions, as well as standard Finnish adopted by originally Swedish-speaking Helsinki residents.
“I’m interested in the effect people’s backgrounds have on language and how quickly language changes at the individual and community level,” Santaharju says.
The Kippo project uses a clustering programme based on Bayesian statistics. It allows the material to be clustered into groups, i.e. speaker populations, based on the frequency of the language variant in question. The clustering programme does not care whether the material consists of genetic or linguistic information, as long as it is fed into the system in a form that the programme is able to process.
The researchers have converted the variants of the language features into numbers according to whether a specific variant is present (1) or not (0). They have also used more complex coding for recording information on how frequently the variant was used.
Another useful aspect of the tool, according to Santaharju, is the fact that it allows the researchers to categorise the speakers without preconceived assumptions. Traditionally, speakers have been classified according to underlying factors, such as age, place of residence or social class. However, these may not be decisive factors. Underlying factors may also have a different effect in different situations. It’s possible that language variation is accounted for by other factors.
People of working age feel the pressure to be well-spoken
The corpus on spoken Finnish collected during the third period is fascinating and even unique. The longitudinal data has made it possible to study not only how individuals’ speech changes with age, but also how language changes from one generation to the next.
The idea that people of different ages speak differently is prevalent in linguistics. On the other hand, it is assumed that once a person is an adult their speech doesn’t change much anymore.
When the material has been collected from people of different ages during a single period, there is no certainty on what the differences in speech mean for language change. It’s possible that the speech of the coming generation, i.e. young people, is a predictor of the direction in which the language is developing. On the other hand, expressions used at a young age may be a kind of temporary stage towards the language used in adulthood.
When the corpus stems from two different times but from the same people, this allows the researchers to observe real changes and distinguish whether the changes occur on an individual or general level. Even then it’s not certain whether the change is permanent or continues to progress – and if it progresses whether it will continue in the same direction.
Material collected at three different points in time offers more reliable data. The corpus of spoken Finnish in Helsinki has demonstrated that an adult’s speech does not become locked in a specific phase and that language variation in an individual’s life does not progress in one direction in an uncomplicated manner. Although people speak differently at middle age than when they were young, their speech may later recover some features of their youth.
“At work, people in their forties clearly feel the pressure to speak in a more standard way. Once they begin to near retirement, they may reintroduce features in their speech that were absent for a time,” Unni Leino explains.
The research reveals facts about the mechanisms of language variation and change in general
In his article published in 2018, Kippo project’s doctoral researcher Olli Kuparinen mentioned that even a short excerpt contains several features that researchers of language variation find interesting. However, in said article, Kuparinen focused his attention specifically on the use of the infinitive in the Finnish spoken in Helsinki.
The Kippo project’s study of the infinitive has provided new information on the progress of its change. It has provided new data on linguistic models of change and how appropriate the different models are in different cases. In addition, the longitudinal data reveals how some language variation and change become established throughout the community.
A special characteristic of the Finnish spoken in Helsinki is the fact that the capital has always been a place that people from all over Finland move to (today, of course, it attracts people from all over the world). The migrants’ dialects all bring something new to the same shared language pot.
“The Finnish spoken in Helsinki is not a single, uniform slang. It has characteristics that are independent of the vocabulary and that allow the speech to be identified as Helsinki-based, but the Southern Häme basis it has been built on is still evident,” Leino explains.
Can the Finnish spoken in Helsinki be used to draw conclusions on the variation and change in the Finnish language on a more general level? Leino explains that this is far from a simple question. Each dialect lives and progresses independent of others, but at the same time, the Finnish spoken in Helsinki influences dialects spoken all over the country through the mass media and the Internet. The perspective of the researchers in the Kippo project is that their research on the Finnish spoken in Helsinki reveals facts not only about the spoken language in Helsinki, but about the mechanisms of language variation and change in general.
Language structures can change quickly too
When a layperson thinks about changes in urban speech, what may come to mind are the slang words used before and now. For Leino, what is fascinating about phonological and morphological variations is the fact that they may reveal information about a deeper change than that occurring at the level of words.
“Vocabulary changes quickly. It reveals changes in culture, the living environment and a person’s employment situation. Because of the Internet, verbal innovations spread much, much quicker than ever before. Phonological and morphological structures are interesting blocks in that variation occurs in them, but they don’t fluctuate in the same way as vocabulary.”
Leino explains that many different types of pressure for change cause structural variation, which is slower. Some changes occur because of a need to simplify the language to make speaking as light and easy as possible. Simultaneously, this is countered by the need to preserve the language in such a state that differences in meaning are retained and speakers understand each other.
Variation in language and pressure to change are inevitable. Standard Finnish as we know it is not the one and only Finnish language; it has a political and historical background.
“Standard contemporary Finnish is a political creation originally constructed in the 19th century by the Helsinki-based educated classes whose primary language was Swedish. For our national ideology, it was necessary for us to have a Finnish language that would serve us in all situations,” Leino says.
According to Leino, Finnish is actually a bundle of extremely varied dialects. There are huge differences between Finnish dialects: they differ from each other significantly even at the structural level. In addition to pure phonetic alternation, there are differences in declination. For example, Professor Lauri Kettunen’s map of dialects presents twelve variants of the verb antavat (they give). Therefore, it is not wise to focus solely on the phonetic level when studying Finnish dialects.
“In the Netherlands, for example, dialectology might be carried out by asking people from different parts of the country to read the same text aloud and then checking what kinds of phonetic variations there are.”
Although words change more quickly than structures, sometimes structural novelties can progress surprisingly swiftly, even from one language to another. Over the last few years, many people have used the sentence construction ‘because x’ where the conjunction is not followed by a complete subordinate clause, but only a noun. I probably won’t see you for a while because COVID.
There are differing opinions on the origin and newness of the structure, but it may have originated in a social media meme that started spreading online in 2011. The meme humorously imitated a writing error or deliberate wording of a classified advertisement for a car on the advertisements website Craigslist: “Completely stripped inside because race car.”
Since then, the same structure has spread to Finland and become a trendy expression – because the Internet.
“It found its place within the language fairly quickly and has been used in ways that don’t directly correspond to the structures already established in the language. That’s why it seems possible – if not likely – that the ‘x’ structure will gradually gain foothold also in standard language and more formal situations,” Leino says.
University lecturer at Tampere University, docent of Finnish and applied computing. Leader of the Kone Foundation-funded project KIPPO – The Temporal Variation of Linguistic Populations 2017–2021.
Evolutionary biologist, postdoc researcher at the University of Helsinki. Specialises in applying the methods of population biology to linguistic corpora; involved in the Kippo and BEDLAN projects.