Rhythm typology seeks to classify languages according to the units on which they base their rhythm. The most used categories are stress-timing, syllable-timing, and mora-timing. Traditional rhythm typology has been questioned by instrumental research, because the claimed isochronous units have not been found. Speech cycling is a method that endeavours to find regular rhythmic units by utilising repeated speech that the speaker produces by accommodating to regularly paced stimuli. This is intended to eliminate the irregularities that would otherwise hide the linguistic rhythm. The present study is a speech cycling experiment on Finnish. The subjects read sentences that were varied in the mora and syllable count. The locations of the stressed syllables were analysed on a relative scale to see if they appear at regular phases. The results show that in Finnish the mora, the syllable, and the quantity pattern all affect the phases.

This paper presents a topic modeling approach to corpus-based dialectometry. Topic models are most often used in text mining to find latent structure in a collection of documents. They are based on the idea that frequently co-occurring words present the same underlying topic. In this study, topic models are used on interview transcriptions containing dialectal speech directly, without any annotations or preselected features. The transcriptions are modeled on complete words, on character n-grams, and after automatical segmentation. Data from three languages, Finnish, Norwegian, and Swiss German, are scrutinized. The proposed method is capable of discovering clear dialectal differences in all three datasets, while reflecting the differences between them. The method provides a significant simplification of the dialectometric workflow, simultaneously saving time and increasing objectivity. Using the method on non-normalized data could also benefit text mining, which is the traditional field of topic modeling.

Finnish nonfinite clauses constitute a complex grammatical class with a seemingly chaotic mix of verbal and nominal properties. Thirteen nonfinite constructions, their selection, control, thematic role assignment, nonfinite agreement, embedded subjects, and syntactic status were targeted for analysis. An analysis is proposed which derives their syntactic and semantic properties by relying on a computational model of human information processing. The model analyzes Finnish nonfinite constructions as truncated clauses with one functional layer above the verb phrase. Research methods from naturalistic cognitive science and computational linguistics are considered as potentially useful tools for linguistics.

This study presents results of two experiments using supervised machine-learning models to examine individual Finnish speakers’ dialectal backgrounds. Data come from interviews conducted with heritage speakers of Finnish in northern Wisconsin and are compared to data from the Finnish Dialect Syntax Archive. The models were constructed and then, following successful validation testing, used to identify the dialectal background of five individual American Finnish speakers. Results showed individual variation in dialectal backgrounds and some correlation to speakers’ likely language input. Our approach offers a new methodological tool for examining speakers’ dialectal backgrounds in situations of language contact.

In this paper, we study natural language paraphrasing from both corpus creation and modeling points of view. We focus in particular on the methodology that allows the extraction of challenging examples of paraphrase pairs in their natural textual context, leading to a dataset potentially more suitable for evaluating the models’ ability to represent meaning, especially in document context, when compared with those gathered using various sentence-level heuristics. To this end, we introduce the Turku Paraphrase Corpus, the first large-scale, fully manually annotated corpus of paraphrases in Finnish. The corpus contains 104,645 manually labeled paraphrase pairs, of which 98% are verified to be true paraphrases, either universally or within their present context. In order to control the diversity of the paraphrase pairs and avoid certain biases easily introduced in automatic candidate extraction, the paraphrases are manually collected from different paraphrase-rich text sources. This allows us to create a challenging dataset including longer and more lexically diverse paraphrases than can be expected from those collected through heuristics. In addition to quality, this also allows us to keep the original document context for each pair, making it possible to study paraphrasing in context. To our knowledge, this is the first paraphrase corpus which provides the original document context for the annotated pairs.

We also study several paraphrase models trained and evaluated on the new data. Our initial paraphrase classification experiments indicate a challenging nature of the dataset when classifying using the detailed labeling scheme used in the corpus annotation, the accuracy substantially lacking behind human performance. However, when evaluating the models on a large scale paraphrase retrieval task on almost 400M candidate sentences, the results are highly encouraging, 29–53% of the pairs being ranked in the top 10 depending on the paraphrase type. The Turku Paraphrase Corpus is available at github.com/TurkuNLP/Turku-paraphrase-corpus as well as through the popular HuggingFace datasets under the CC-BY-SA license.

This article describes variation in the use of frames of reference (FoRs; object-centred, viewpoint-centred, and geocentric, as in Holistic Spatial Semantics) in Finnish descriptions of motion and connects questions of variation to a typological framework. Recent research has described the choice of FoRs as a process with multiple factors. This complexity and controlling for the main variables posited in the literature create the starting point for the current study that explores factors affecting the choice of FoRs in motion situations and within speakers of the same language. The data were elicited from 50 native speakers of Finnish by using video stimuli. The informants were (mostly) formally educated young adults living in urban surroundings. The analysis reveals considerable variation in individual coding strategies, especially in the inclusion of the speaker’s viewpoint. It also considers variation with respect to different types of trajectories and cross-linguistic differences in the resources of spatial reference.

Sounds spoken on the inbreath have been shown to be common in the world’s languages, and in the Nordic languages ingressive speech seems to be especially frequent. The present study focuses on Finnish and Danish response particles spoken on the inbreath, by examining their uses in everyday talk-in-interaction in corpora of recorded interactions. The particles we examine and their non-ingressive counterparts can perform confirming and acknowledging actions. We analyze the particles as receipts to answers to questions, as responses to questions, as responses to assessments, and as responses to affiliation-seeking utterances. In these positions, the ingressive particles turn out to index that the content of the previous turn was already sufficiently established and, consequently, that there is nothing to add. In cases where an engaged response is called for, the particles are shown to have a disaffiliative potential.

In this paper, I examine a case of vowel insertion found in Savo and Pohjanmaa dialects of Finnish that is typically called “epenthesis”, but which demonstrates characteristics of both phonetic excrescence and phonological epenthesis. Based on a phonological analysis paired with an acoustic corpus study, I argue that Finnish vowel insertion is the mixed result of phonetic excrescence and the phonologization of these vowels, and is related to second-mora lengthening, another dialectal phenomenon. I propose a gestural model of second-mora lengthening that would generate vowel insertion in its original phonetic state. The link to second-mora lengthening provides a unified account that addresses both the dialectal and phonological distribution of the phenomenon, which have not been linked in previous literature.

This article examines evidentiality in the frame of inferential adverbs in written interaction from the perspective of Finnish, a language that does not have evidentiality as a grammatical category. The analysis focuses on six adverbs, such as käsittääkseni ‘as far as I understand’ and tietääkseni ‘to my knowledge, as far as I know’. Evidentiality and epistemic modality intertwine in their semantics, as these adverbs represent a writer’s access to information, but also indicate her evaluation of its reliability. First, this article offers a description of the interactional functions of these adverbs such as marking a writer’s opinion in contrasts, expressing slight hedging in order to anticipate corrections, to allow space for other opinions, or to create irony. Second, in the framework of cognitive grammar, the analysis focuses on the meaning of the evidential verb+kseni construction and the effect of different verb stems on it. These adverbs share similar functions in texts, which is due to their flexible constructional meaning. While varying from lexeme to lexeme, specific evidential and epistemic dimensions can either be foregrounded and relevant in a situation or remain backgrounded and not activated.

Finnish word order is relatively free, making room for all mathematically possible word orders in many constructions. Because there is no evidence in this language for radical nonconfigurationality, explanations must be sought from syntax. It is argued in this article that morphosyntax and word order represent syntactic structure at the PF-interface. Rich morphosyntax frees word order, poor morphosyntax freezes it. The hypothesis is formalized within the context of a parsing-oriented theory of the human language faculty (UG) combining left-to-right minimalism with the dynamic syntax approach. The analysis was implemented as an algorithm and successfully tested with a corpus of 119,800 unique Finnish word orders.

This article examines the Finnish tail construction (right dislocation) used as a first mention of a referent and the variation of the demonstrative pronouns tämä ‘this’, tuo ‘that’, and se ‘it’ in the construction. Many previous studies have suggested that tail construction (TC) referents are highly active and thus already mentioned and salient in a conversation. However, in Finnish, the TC may introduce new referents into a conversation, and this article provides an empirical analysis of how and why this is done. First-mention TCs are often evaluations or questions in which the proposition links the utterance to the preceding context. When presenting new information, the TC allows the speaker to present a potentially lengthy lexical definition of a new referent at the end of the utterance, avoiding the additional emphatic meanings or unwanted implications a simply inverted word order might create.

It has been suggested that second languages and translated languages are constrained by an interplay of several linguistic systems. This paper reports on a data-driven quantitative study on constrained Finnish. We detect linguistic phenomena that distinguish constrained from non-constrained Finnish across constrained varieties, first/source languages, and registers. Implementing a two-phase method, we first detect key quantitative differences of syntactically defined POS bigrams between each variety-, language-pair- and register-specific constrained dataset and its non-constrained counterpart, using Boruta feature selection. We then use the results as variables in a Multi-dimensional Analysis. The results show that both nominal complexity and verbal/clausal complexity distinguish constrained from non-constrained Finnish. These differences interact with both type of constraint and register: the constrained varieties are less sensitive to register differences, and this tendency is more pronounced in learner Finnish than in translated Finnish. Leaving out any of these variables from the analysis would blur our view of this multi-faceted phenomenon.

In this article, we examine the form and meaning of a subtype of causative adjuncts, purpose adjuncts, in Estonian and Finnish. The analysis is based on the micro-modular approach of Conceptual Semantics (Nikanne 1990, 2005, 2018; Pörn 2004; Paulsen 2011a, 2011b; Petrova 2011). It provides a non-linear approach to linguistic causation, alternative to the linear models. By the term ‘non-linear’ we mean the involvement of separate levels of conceptual structure that influence the causative reading, and a separation of causation from the temporal flow of the situation. We propose two causative operators for subordinate causative structures, BECAUSE OF and LEAD TO. According to their temporal structures, we separate two types of purpose adjuncts: reason-purpose adjuncts and avoidance adjuncts. The causative effect of purpose is in our analysis related to the notion of intentionality.

Most English compounds are spaced compounds, whereas spelling regulations prescribe Finnish compounds to be written in a concatenated format. However, as in English, Finnish compounds are commonly spaced nowadays (e.g., piha juhla ‘garden party’), a phenomenon that we labeled the ‘English disease’. In this eye movement study with Finnish–English bilinguals we investigate whether the reading of a concatenated or illegally spaced Finnish compound is affected by the spelling of an English translation equivalent (ETE). We found that spaced Finnish compounds were read slower than their concatenated counterparts, but this effect was attenuated when ETEs were thought to be spaced. Similarly, concatenated Finnish compounds were read faster when their ETEs were also concatenated. These backward transfer effects are in line with studies that show that processing behavior in L1 is affected by a strong concurrent L2, even when the L1 is the native language as well as the dominant community language.

This paper deals with the grammatical differences and overlaps between the uses of the Finnish kuulua as a verb of auditory perceptibility (‘to be audible’) and as a verb of appearance, when employed in negative clauses (‘to be imperceptible (through unspecified sensory input)’). Both meanings entail perceptibility, existentiality and motion from the experienced towards the experiencer. However, they differ significantly in regard to the nature of the motion as well as the degree of animacy of the subject referent. As a verb of auditory perceptibility, kuulua accepts mainly inanimate subjects referring to a perceivable sound. As a verb of appearance, kuulua is mostly used with animate subjects. The semantic difference between the two constructions is accounted for in terms of objective and intersubjective meaning construal. The potential movement of a sound towards the experiencer concerns the relationship between the world and the subject of conceptualization, whereas the non-appearance of an animate being is viewed on the level of intersubjective cognitive coordination, with regard to interactional expectations. The results of this study shed light on the complex semantics of perceptibility. The analysis is based on 1,528 occurrences of kuulua in dialectal and literary data.

The distance or similarity between two languages can be objective or actual, i.e. discoverable by the tools and methods of linguists, or perceived by users of the languages. In this article two methods, the Levenshtein Distance (LD), which purports to measure the objective distance, and the Index of Perceived Similarity (IPS), which quantifies language users’ perceptions, are compared. The data are the quantitative results of a test measuring conscious perceptions of similarity between Estonian and Finnish inflectional morphology by Finnish and Estonian native speakers (‘Finns’ and ‘Estonians’) with no knowledge of and exposure to the other (‘target’) language. The results show that Finns see more similarity between Finnish and Estonian than Estonians do. Also the correlations between LD and the perception results of the Finns are statistically significant while the correlations between the LD and the IPS scores of the Estonians are not. Comments by test participants provide insights into the nature of the perceptions of similarity.

Several factors affect the comprehension of a text written in a language related to the reader's first language (L1): (i) orthographic similarity with the reader's L1, (ii) contextual clues, (iii) semantic relationships between components of phraseological units, and (iv) L1 reading comprehension strategies. This article compares the results of a cloze test (CT), in which a group of Finns read a text in their L1 (Finnish) and filled the gaps, and a translation test (TT), in which another group of Finns translated the Estonian version of the same text into Finnish. This text included five pairs of primes and targets, parts of the same phraseological unit, representing different semantic relations; in the CT the target was replaced by a gap and in the TT the respondents had to translate the target. The results indicated that the respondents used similar inferencing strategies in both tests, and thus provide evidence for the assumption that orthographic similarity is not the sole factor contributing to understanding a foreign text, but that L1 reading comprehension strategies are also employed.

Among the characteristic features of the Finnish language is the use of numerous derivational affixes and diverse word-formation options. Although Finnish has very old derivational elements, fairly recent suffixes and even completely new ways of forming words are also found. It is typical of word-formation options that they change, and that their frequency and popularity varies over time. In this diachronic study, the focus is on one of the most recent suffixes used in the Finnish language, the agentive -Ari suffix (e.g. kaahari ‘reckless driver’, kuohari ‘gelder of animals’). What makes the -Ari derivatives special is that the type has been adopted on the model of words borrowed from the Germanic languages. Historically these are descended from the Latin derivational element -ārius, which was adopted widely in the European languages. The main purpose of the present study was to find out whether, from a diachronic perspective, the -Ari-derived agent nouns actually represent an independent derived semantic category in Finnish. Another purpose was to characterize the process whereby the -Ari suffix was adopted in Finnish: at what point do these derived forms actually first occur in Finnish, and how has the use of the derivational element been manifested at different times. A final significant task of the study was to clarify the potential reasons and motivations for this morphological borrowing.

This article compares three accounts of Finnish intonation using a perception experiment with manipulated f0 contours. The experiment involved compound/noun phrase minimal pairs differing in f0 pattern. To address the question of tonal specification, manipulations changed f0 contours of recorded compound words, associating them with f0 patterns having different components of the naturally occurring f0 rise-fall contour. Thus, the study investigated which tonal targets were crucial for the perception of a complete tonal contour inducing a noun phrase interpretation. Results suggested that the falling part of the rise-falls, modelled as realisations of a high and a following low target, was essential. They furthermore revealed evidence for these targets being associated with prosodic phrases, as well as for Finnish tonal targets being characterised by a flexibility that contrasts with accent realisations in languages like English.