Register features in Kalevala-metric incantations

Genres and registers of Kalevala-metric poetry have customarily been treated as correlating concepts. The paper examines a category of Kalevala-metric Origin of Fire incantations as a text-type which includes multiple linguistic registers functioning inside the genre. These registers are reflexive formations, potentially established as social situations when addressing supernatural agents. The analysis of directive utterances in the data reveals that addressing of agents varies according to the addressee. An examination of moderative particles and justification of performer's actions results in a view of registeremblematic means of addressing different agents. Register-based approach provides a way to critically re-asses the relationship of genre and register in Kalevala-metric poetry. 2021 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

and hostile agents in incantations have the potential to be distinguished as different types of social situations and therefore to employ different registers. The present paper sets out to demonstrate such distinctions and possibilities of utilizing the concept of reflexive registers in research on oral poetry. I will review a dataset in order to investigate if distinctive registers can be identified functioning in a singular incantation type. My hypothesis is that differently evaluated non-human agents are addressed in different registers. In other words, this would mean supporting agents of a ritual specialist would be addressed differently than hostile actors.
Methodologically, I will analyze the variation in address by reviewing directive verbs, particle use, the optative mood, and justification present in the texts. The analyzed particles -pa, -pä, and -s, as well as the optative mood can grammatically index moderation in standard language (Ahtia, 2014, 87;also Hakulinen et al., 2004, x 835, x 838). The role of the optative mood is to employ much weaker directive force than the imperative (Kangasniemi, 1992, 98;Leskinen, 1970, 66). I hypothesize that the particles investigated in the analysis moderate the directive force of imperative verbs in Kalevala-metric poetry. I further hypothesize that these and the optative mood are predictably used in incantations according to the attitude or relationship of the performer to the addressee.
The research questions are therefore the following: 1) By what attributes are supporting and hostile agents distinguished from each other? 2) Are directive verbs with particles or the optative mood and justification used as indices of moderation in the data? And finally: 3) Do directives and particles operate mostly to meet metrical needs by filling a position where a syllable is required, or do they potentially index different relations between the performer and the addressed agents? By answering these questions, I hope to propose a more detailed view concerning workings of discursive registers in an oral tradition (cf. Gintsburg, 2018, 106).

Research material
The data analyzed in the present study is taken from the published edition of approximately 87,000 Kalevala-metric texts, Suomen Kansan Vanhat Runot, or SKVR (Ancient Poems of the Finnish People) 1997), available in a digital edition that forms an electronically searchable corpus (skvr.fi). The anthology was originally organized first by geographical region, and then texts from most regions were organized according to a series of ideal categories that correspond roughly to etic genres and ideal text types within genres. As emphasized above, the etic category of incantations includes a wide variety of materials. These materials cannot all be assumed to form a single emic genre. These problems are addressed here by narrowing the study first by regional focus and second by a particular text-type category of incantations identified as The Origin of Fire (Tulen synty). The regional focus on traditions of Viena Karelia (see map 1) reduces the potential heterogeneity in the emic genre models encompassed by the data. Origin of Fire incantations form a narrow category that meets three criteria: (a) they are communicative incantations, as defined in Ch. 3; (b) the category is distinguished by the manipulation of cosmogonic knowledge, which allows incantations of this type to be strongly linked to the ritual specialist as a prototypical user; and (c) incantations of this type are generally connected to healing injuries connected with fire. Although Origin of Fire incantations were likely conceived as belonging to an emic genre that also included other types of incantations, this narrowing of the data to incantations of a single type allows the texts analyzed to be identified with a single genre without becoming entangled in questions of what other incantations did or did not belong to the same genre.
The resulting data is constituted of 46 texts of which the earliest is documented in 1829 and the latest in 1915. These texts range in length from 22 to 216 lines with a total of 3449 lines in the analysis.
The methodological problems with the data are mainly related to ethnographic data documentation situations and archival processes. The documented material has been collected by means of interviews, of which many were dictated and handwritten as notes, and contextual information about the interview situations is rather scarce. Furthermore, especially in the early period of data documentation (approximately the first half of nineteenth century), the prevailing collection paradigm deemed names of informants and their possible status as ritual specialist secondary. These factors make it hard to determine if an atypical text or line is an intentional variation, is simply idiolectal, is owing to a clumsy delivery or lack of knowledge, reflects a gap in the performer's memory, mishearing or some other mistake by the interviewer, or any of a myriad of other possibilities. Owing to these issues, the relevance of working with this old, extensively researched source material has even been called into question (cf. Wolf-Knuts, 2020, 116), a trend that climaxed in Finnish Folklore Studies in the 1970s with the emergence of the performance-oriented paradigm (e.g. Wolf-Knuts, 2020, 118). Fortunately, the value of large archival collections is recognized today, although the material must be approached through multifaceted source criticism (see e.g. Kurki, 2004;Wolf-Knuts, 2020 for more recent viewpoints; on the political nature of folklore collection, see Anttonen, 2005).
The research I conduct in this paper takes as a point of departure the assumption that the so-called authenticity or factual reliability of archived texts does not prevent the possibility of applying varying styles of reading. I focus on the prospect of the material answering on the relevant questions posed (Kalela, 2000, 92.). For this purpose, the incantation material forms a suitable corpus, recognizing that my interpretations as a researcher become juxtaposed with the vernacular views (Tarkka, 1993, 167).

Kalevala-metric incantations
Kalevala-metric poetry describes a common oral-poetic system that was shared by the Finns, Karelians, Ingrians, Estonians, Setos, Votes and Ingrian-Finns as linguistically related ethnic groups (see Kuusi et al., 1977, 34-37; for discussion of the term Kalevala-meter, see Kallio, 2011, 391;Kallio et al., 2017). In terms of poetic structure, Kalevala-meter is a trochaic tetrameter, thus verses normally have eight syllables, and it includes rules governing the placement of long and short stressed syllables. The poetry is also characterized by verse-internal alliteration, though not required in every line, parallelism and what can be called right justification (i.e. placing longer words at the end of the line) (Frog and Stepanova, 2011;Kuusi et al., 1977, 62-68;Leino, 1986, 129-146;Kallio et al., 2017.). The excerpt below demonstrates some of the main features of the meter, perhaps most prominently alliteration (e.g. siipesi/simaha) and parallelism (tuo simoa siivessäsi/mettä kielesi nenässä): (1) Meheläini ilman lintu Honeybee, bird of the air Kassa siipesi simaha Dip your wing in mead Sulkasi sulaha voihe Your feather in liquid butter Tuo simoa siivessäsi Bring mead in your wing Mettä kielesi nenässä Nectar on the tip of your tongue (SKVR I 4 , The oral poetry was documented especially during the nineteenth and early twentieth century and the traditions varied considerably among different groups. The present study focuses on communities of the region Viena Karelia, also known as White Sea Karelia, White Karelia or Archangelsk Karelia (see map 1), where vernacular mythology and the associated incantation tradition were prominently found during the nineteenth century, making it a major destination during the enthusiastic documentation efforts in the era of National Romanticism.
Socio-historically, in the time of major oral poetry documentation, the Viena Karelian area was characterized by low literacy, syncretic vernacular religion and lack of administrative presence (Tarkka, 2016a, 176). The subarctic climate and land division practices affected farming, leading to a mixed subsistence economy, which also included itinerant cross-border trading (Tarkka, 2013, 30-33). In the time of major oral tradition documentation, Kalevala-metric poetry served not only as an aesthetic means of communication of salient information, but also as a medium for local issues and personal emotions (Tarkka, 2016a, 176).
Kalevala-metric incantations are very heterogeneous: they have most often been treated as a general category for all texts in the poetic form characterized by a function of changing surrounding reality (Tarkka, 2013, 68, 67-72; see also Anttonen, 1999, 19-21;Kallio, 2013;Kuusi, 1963;Tarkka, 2016a;Timonen, 2004). One way to approach the diverse group of incantations has been a division into those that anyone could know and use, and those used only by ritual specialists. In agrarian communities, the commonplace incantations were generally short and understood as mechanically affecting the world through accurate recitation of a text with any required accompanying actions; such incantations were widely used to guarantee good fortune in fishing, cattle-luck, and in other practical matters (Siikala, 2002, 76). Incantations used by specialists called tietäjät (sing. tietäjä; literally 'one who knows, knower') functioned on other principles. The specialist incantations relied on ritual-specific communication with supranormal agents. They differ from mechanically effective incantations through reliance on personal knowledge of the supranormal otherworld, the innate power of the performer, strategies for communication with agents of that world, and associated rite techniques (Siikala, 2002, 77-79, Tarkka, 2013. Researchers have often divided Kalevala-metric incantations into categories such as prayer incantation (rukousloitsu) and incantations dealing with origins of a phenomenon in the world (syntyloitsu) (Siikala, 2002, 84). Prayer incantations get their name from the invocation of the Christian God, Jesus, Virgin Mary, or similar agent as a part of the incantation, while incantations dealing with origins include a presentation of the mythic origins of fire or other phenomena, thus presenting a vernacular incantation ideology that knowledge of the origins of a phenomenon yields control over it (Siikala, 2002, 86). The origin incantations were generally used with the aim that was most common for incantation use: healing (Siikala, 2002, 85). With access to mythic knowledge concerning the origins of a phenomenon in question, the performer of the incantation was in theory able to heal various injuries.
A problem with these categorizations is that they are based on a single prominent feature of form or content although a single text in the corpus may include several such features (Siikala, 2002, 84-85). In the context of this study, a more useful categorical distinction is made between communicative and non-communicative incantations, according to which the salient and uniting feature of communicative incantations is the address of supranormal actors (Siikala, 2002, 84-92).
Example (1) above is from a communicative incantation and illustrates the direct address of a supernormal agent, the bee (mehiläini) who moves between worlds when directly addressed by the tietäjä. The bee of the incantation is not any ordinary creature, but a mythic animal helper of the ritual specialist (see Siikala, 2002, Part V). The following text illustrates the nature of non-communicative incantations. The incantation, aiming at a successful outcome of a trial, lacks a communicative aspect of contacting supernatural agents at the linguistic level: ( Instead of seeking to contact supernatural agents for support, the speaker of the incantation above trusts in the correct form of the incantation and their willpower to guarantee an agreeable result. It should be noted, though, that even the communicative incantations could include non-communicative passages.

Genre and register
I handle genre and register as complementary concepts. I treat genre here as a set of prototypical features, which guide both the production and reception of texts of a socially recognizable type in situated performances (see Hanks, 1987, 870). In other words, I approach genre as an orienting framework for the production and reception of particular types of text which are produced and interpreted by their intertextual links to earlier texts (Bauman, 2000, 84;2004, 3-4). The fundamental feature of genre is then its indexicality to other texts of a common type (Visakko, 2015, 43). Genres are nevertheless social formations that are linked to and evolve through conventional users and uses (Visakko, 2015, 44). Consequently, the indexicality of genre extends from the formal dimensions of texts of the same type to the social dimensions of how, when, why, and by whom that type of text rather than another is used in society, as well as the types of conventions or potential for mixing of prototypical features linked to other genres (Frog, 2016b;Tarkka, 2013, especially 88-102). Where a genre is prototypically performed by a particular social identity, it may also index that identity while features of texts of that genre are historically shaped by the perspectives of that identity's social position and stance-taking in performance (Frog, 2015, 88). For genres of ritual poetry, this commonly means that the genre is characterized by the perspective of a ritual specialist as the prototypical performer.
Folklore genres may be complex in the sense that genres or parts of genres may be embedded within a single text (Bakhtin, 1986) or the positioning of the performer may vary during the course of a text rather than remaining unilateral and uniform. The question of distinguishing a genre or genres of Kalevala-metric incantations will be returned to when presenting the data for this study. Here, it is sufficient to offer the following example of a Kalevala-metric incantation from that data. The example naturally divides into five parts. The incantation opens with a statement of knowledge of the origins of fire in mythic time. Following the rather common style of presentation, such knowledge is presented as third person epic narration (Part I). The performer then describes a contemporary mythic agent in the unseen world (Part II). The performer addresses this agent in direct speech (Part III). The performer then addresses a different agent in direct speech, in this case in order to ensure success (Part IV). The incantation is concluded with soothing verses of healing (Part V). (3) Example of an The weak one could draw his breath Yön tullen pimiten päivän When the night comes and day darkens 55 Ainap on auringoa aleten Always when the sun goes down Jotta sais sairas maata So that the sick one could-s lie Heikko hengensä veteä The weak one could breathe (SKVR I 4 , 294.17-57.) 1 The -p at the end of tihop is a metrically-motivated construction used as a metrical filler. 2 The -s here, as well as in further occurrences in the paper, demonstrates suffix-usage in directives: viskoa (toss imp ) vs viskoas, where the suffix can indicate moderation of imperative utterance. 3 The on (is) is another metrical filler.
Here, the concept of register is adapted from its use in linguistic anthropology. Although a register can encompass both linguistic and non-linguistic signs (Agha, 2004(Agha, , 2007, the nature of the data in this study limits consideration to linguistic signs, i.e. the linguistic register. A linguistic register is similar to a genre in that it is constituted of a variety of usable linguistic signs and grammatical forms (Agha, 2007, 169) that index stereotypical social identities (such as gender and profession) or situations of interaction (Agha, 2007, 14, 55;Blommaert, 2007, 117;Lehtonen, 2015, 36;Visakko, 2015, 46). In the case of folklore genres, the stereotypical identity indexed by a register or registers may also be a text type category (Frog, 2015, 92), linked to ideas of the texts as "things made of language". This means that the same register may be used for verbalizing the text in spite of variation in the social identities of individual performers and specific situations of performance or dictation (Frog, 2019). Because folklore genres may be complex, multiple registers may be employed in the production of a text of a single genre. Many incantations open with third person epic narration that may be specifically linked to the incantation tradition or might also be performed as part of a whole epic narrative. Within Kalevala-meter poetry, certain formulae and verse sequences index particular genres and these uses of epic in incantation are recognized as involving shifts between registers linked to different genres (Tarkka, 2016b; see also Frog, 2019, 227). In example (3) above, the performer addresses two different otherworld agents, and in some incantations the performer may take an aggressive or hostile stance, for example banishing a cause of illness as an adversary (Haavio, 2019(Haavio, [1967, 331-341). If registers are viewed as not only indexing stereotypical social identities but also situations of interaction, the data in this study can potentially be used to explore whether incantations may also include shifts in register according to the relationship between the speaker and the addressee.

Methodological considerations
In the following analysis, I review all directive verbs and use of particles quantitatively. The purpose is to consider a possible differentiation in their usage between requests and commands. Following that, I perform an analysis of characterizations of addressees, which aims at defining agents as either sympathetic or hostile to the speaker. The inspection of the data is performed after that to reveal a possible correlation between agents and directive utterances, verbs of action and movement, and an assessment of metrical requirements related to the lexical forms used. In addition, I shall look at how additional features co-vary with positive agents. An overview of features is then provided.
One methodological issue in the quantitative analysis is that the relative frequency of a feature is affected by differences in usage found in individual texts and the length of the respective incantations. The statistical distribution of the directives with particles shows that 67% of affixed forms occur in relation to directives aimed at positively evaluated actors, while 28% are used in connection with ambivalent agents, and 5% with negative tokens. Some of the performers employ greater numbers of parallel verses, where parallel verbs are customarily in the same mood (e.g. lines 50-51 in example (3)), thus greatly increasing the occurrence of the form in the data (see e.g. SKVR I 4 , 285). In contrast, some of the incantation texts include only one or two directives (e.g. SKVR I 4 , 251). The differences in the number of directives can be due to the idiolects of the performers, but the documentation situation and collector are additional factors to be considered. While some collectors made very attentive linguistic transcripts, others could omit words or whole verses in their notes (see e.g. SKVR I 4 , 325).

A hypothesis of moderation in directive utterances
Directives are modal uses of verbs that ask, urge, command, or advise the receiver to perform an action, and these modal uses may vary in strength (Lauranto, 2014, 12;Sorjonen, 2001, 93;Hakulinen et al., 2004Hakulinen et al., x 1645. The data contains 170 verbs used as directives. Of these, 39% affix the particle -pa, -pä, -pas, -päs or -s to the verb or inflect the verb in the optative mood (i.e. with -os, ös, -as, äs). Although the particles -pa and -pä are distinguished in analysis, they are variations of a single morpheme: Karelian is characterized by what is called vowel harmony and this morpheme variation is determined by the vowels in the preceding word stem. The vowel-harmony distinction of suffixes in/a/and/ä/was made to ensure that prominent use of the particle with a particular verb like mennä ('go') did not lead one form of the particle to develop distinctive patterns of use rather than remaining a mere allomorph.
In Kalevala-metric poetry, the specific metrical requirements of the poetic mode complicate the question of modal significance. Especially with two of the longer moderative suffixes, the effect is twofold. While the occurrence of -pa and -pä can indeed function as moderative of the utterance, they can also act in the metrical environment as fillers to achieve the required minimum number of eight syllables, and potentially to affect the positions of subsequent syllables owing to metrical restrictions on the placement of stressed syllables according to their quantity.

Actor types as addressees
In order to assess the possibility of a difference in register reflecting a difference in social relations between the performer of an incantation and the addressee, it is necessary to distinguish mythic actors being addressed according to different types and their potential relations. Actors addressed in the data have been assessed centrally according to three criteria: (a) whether the agent has an identity generally established in the genre or corpus; (b) whether there is an evaluation implicit in the context of interaction or description; and (c) whether there explicit or implicit evaluation is observable in epithets, parallelism or other verbal references to the agent. Evaluations are assessed as positive, negative or ambivalent in relation to the performer. The three categories of assessment are considered complementary. Of particular interest is whether there is a correlation of contextual and verbal evidence that points to epithets or other phraseology as characteristic of distinct registers rather than being specific to the particular agents mentioned.
Some of these actors, such as the Virgin Mary, have identities that are developed through a wide range of discourse that extends well beyond the scope in the Origin of Fire incantations and beyond incantations more generally, while other actors, such as the honeybee (mehiläini), are more specifically identifiable with incantations and communicative incantations in particular. Contextual evaluation as positive or negative is based on how the agent functions relative to the performer's goals in the incantation (healing the patient). If the agent's function supports those goals through positive action, the agent is considered positive; if the performer seeks to inhibit the agent's actions or expel the agent entirely, that agent is considered negative. These evaluations may also be reflected in descriptions where the speaker does not engage in direct contact with the agent. These evaluations are correlated with implicit or explicit evaluations in the language and phraseology of the incantation. Such evaluations typically occur in verses that contain no directives but endow actors with, for example, positive attributes: Neit syt Moarie emo ni Virgin Mary mother diminutive Raka s äiti armolli ni Beloved merciful mother Se on va nhin vaimok siista She is the oldest of womenfolk Eläjiistä ensimäi ni The first of the living (SKVR I 4, 287.1-4.) The most frequently used positive attributes are raka s äiti armolli ni ('beloved merciful mother') with 16 tokens, jumalan luoma ('God's creation') with 23 tokens and emonen ('mother diminutive ') with 25 tokens. The most numerous of the epithets, emonen, occurs almost solely with Virgin Mary. While some of the texts also include other female otherworld agents (such as Icy Maiden above), they are generally addressed by their (e.g. ice-related) attributes. The word emonen, however, also appears as a noun instead of an epithet in some cases where the performer is threatening to seek out the mother of a harming substance, e.g. fire: (10) Tules työsi tuntomaha Come-s to know your work Pahojas' parentamaha To heal your evils Ennenkuin etsin emon Before I look for the mother Nouvan valtavanhempas Fetch your forebears (SKVR I 4 , 293.34-37.) The third most common of the attributes, raka s äiti armolli ni, is solely associated with Virgin Mary in this data. The epithet jumalan luoma in turn is used only with agents of subordinate status, such as the honeybee. Most commonly jumalan luoma occurs in verses where fire is invoked or commanded, with 15 tokens.
Evaluation may also take the form of deixis and contrast. In the data, the honeybee is widely valorized as being an agent that brings healing ointments at the order of the performer, and is generally in the role of helper. The honeybee may be contrasted with the hornet, associated instead with venom or threatening potential: (11) Meheläin on meijän lintu Honeybee is our bird Hörhöläin on heijän lintu Hornet is their bird Lennäpä tuonne kunne käsken Fly-pä there, where I command (SKVR I 4 , Here, the primary verse identifies the honeybee as "ours" paralleled by the hornet as "theirs". In contrast, a different performer may, rather than identifying parallelism of contrast, describe this agent through parallelism of equivalence or identity: (12) Herheläi ni Hiien lintu 4 Hornet, bird of Hiisi Mehiläini ilman lintu Honeybee, bird of the air Lennäs tuolta liuottele Fly-s from there, hasten (SKVR I 4 , This kind variation between contrasting and identifying actors occurs in three incantations, and it is difficult to judge whether they constitute examples of counter-valorizations or a slip of the performer's tongue or memory. The main point is to acknowledge that ideological metapragmatic evaluations can, at least in theory, have counter examples alongside the evaluations that otherwise appear dominant in the corpus. The second category consists of agents whose status is ambivalent: they are either lacking a clear position in the positive/ negative dichotomy, or their status varies contextually. The most prominent example in the data is fireits status looms between positive and negative both within and between texts, while the connecting feature is fire's power. As mentioned above, fire is frequently addressed as "God's creation", which indexes a positive status and commensurability to humans. In the data, however, fire is also an injury-causing agent. This ambivalent status is manifest in the way fire is addressed. The total number of directives in the data addressed to fire is 28. Of these, 39% appear to employ moderation by particle-use or follow the directive with a justification, while the rest of the tokens are simple imperatives. Even when fire is addressed with imperative directives, the tone is generally neutral and employs reflexive verbs: (13) Tuloni Jumalan luoma Fire diminutive God's creation Tulihisi tummennaite Mar yourself in your blaze Porohisi peitteleite Cover yourself in your ashes (SKVR I 4 , There are, however, signs of a different kind of register in three tokens. These include lines such as otas viinana vihasi ('take-s your wrath as booze') and tules työs tuntemahan, pahasi parantamahan ('come-s to know your deeds, to cure your evils') (SKVR I 4 ,330;I 4 ,262). The lines are formulae typically used for banishment in incantations on other themes (see Siikala, 2002, 92). Their rarity in the data suggests that, while fire is in the opposing position to the speaker, it does not require the same level of aggression as some other agents. The three tokens above are the only examples of treating fire as clearly negative in the data.

Justification as a means of moderation
By justification, I refer to verses directed to an agent and presenting a reason for the speaker's request (see below). In the data, there are eight tokens of justification, six with directives employing particles and another two with simple imperatives. I have extracted the tokens from the data by examining the speaker's direction of address in the verses surrounding directives. The form of the justification does not exhibit variation between the uses with imperatives and with verb forms using a particle. The tokens of justification follow a similar pattern throughout: the performer of the incantation issues a directive to an actor, either as a direct imperative or formed with a particle. The directive is then followed in subsequent verses with a statement of the reason for the request or command: (14) Neitsyt We can see that the directive issued to the Virgin Mary is followed by a justification, where the reason for the call for help is explained as avoiding or lessening the amount of pain and securing the injury from infection. Justification, however, does not occur solely with directives to positively evaluated agents. In two tokens (SKVR I 4 ,295;330), the performer justifies their directive to a negatively evaluated agent. In the first of these cases, imperative lines to fire connect with an argument for the patient's sake: (15) Tuli nouse taivosehe Fire, rise to heaven Auver pilvihi ajate Smoke, flee to clouds Ihosta imehnis-raukan From the poor human's skin Emon tuoman ruumehista From the body of one brought by a mother Jotta saisi sairas moata So that the sick may lie Heikko henkeñsä levätä The weak rest his / her spirit (SKVR I 4 , The second occurrence of a combination of a negative agent and a justification incorporates the ascription of positive attributes to the agent: (16) Tulo ne Jumalan luoma Fire, God's creation Panu Jesuksen panema The blaze set by Jesus

Otas viinana vihasi
Take-s your wrath as booze

Metenä mielimuurumasi
Your grief as nectar Ettei tuskilla tulisi So that it would not bring pain Eikä karstoin katkiesi Nor break off with soot (SKVR I 4 330. 38-43.) The combination of formulaic banishing verses and justification is unusual, perhaps doubly so since the speaker combines use of an s-particle with the directive. This may be either the result of the performer's idiolect (the incantation includes frequent use of the s-particle with other agents as well) or based on the view of fire as God's creation, which requires moderation. All things considered, the inclusion of a justification correlates strongly with positive attributes of agents. This seems to indicate a role of justification as a rhetorical device emblematic in a moderative register. The number of examples nevertheless remains relatively small and a fuller investigation is needed.
9. The particles -pa and -s

The data
This section introduces the range of verbs and lines where the particles occur. I have focused solely on directive utterances, which means that I have excluded, e.g., declaratives from the analysis, as the particles do not occur in them. This enables focusing on a verification attempt of the main hypothesis. Below, I present the directive verbs categorized according to mode of action (cf. Saarinen, 2018, 197-209).
The first group consists of transitive verbs describing activity that results in the movement of an object from location A to B. The objects in this group are healing substances (nectar, frost, ointments) or artefacts (a mitten, bath whisk) used by the Roll-s knife from heaven -s 2 performer in the healing incantation. The subjects performing the movement in this group are implicitly present in the verses and are described as mythical helpers, such as the Virgin Mary or the honeybee. I provide the lines with translations in Table 1. The second group demonstrates verbs of movement. The lines of this group do not include movement of substances, but the aim is of achieving transformation of actors, shown in Table 2.
The third group of verbs includes verbs of action, of which some feature an object and some do not, shown in Table 3. Table 4 lists the verbs with their particles and imperative forms (see Table 5 for total particle numbers).

Particles in context
The particle -pa occurs in three variations in the data, -pa, -pä, and with apocope as -p' on. The most numerous type is -pa, with five tokens. Tokens of this form appear for example in following verses: (4) Tuopa sie mettä siipesi sisässä Bring-pa nectar inside your wing Tahi kielesi nenässä Or on the tip of your tongue (SKVR I 4 , Functionally identical to the -pa type is -pä. The difference between -pa and -pä is based on vowel harmony introduced in x5 above and is a matter of phonological complementary distribution (-pä appears where different phoneme is required):  As we can see, in this type, the particle is simply added to the verb form. A variation of this type is one where the final vowel of the particle is omitted and the verb on ('is') is added: (17) Tuop' on noita voitimija Bring-p' (is) those ointments Joilla Luoja voijettih With which the Creator was anointed (SKVR I 4 , The data does not include any examples of the shortened for -p' without being followed by on. The number of this type in the data is four tokens. In addition, I have included in this particle group a type with two tokens, -(p)pas, which is a -pa þ -s suffix; there are no examples of this suffix with the vowel/ä/in the present data: (18) Tuoppas lampi tullesasi Bring-pas a pond with you Jolla tulta tuimentelen With which I make fire bleak (SKVR I 4 , The total number of tokens with the particles -pa and -pä and their s-variants is therefore thirteen.

The optative mood
The optative mood in the data occurs in second person directives, which request an actor to perform an action. The total of optatives in the data is 25: Tuuvos mettä tulle s sasi Bring-os nectar when you come Simoa kielezi nenä s se Mead on the tip of your tongue (SKVR I 4 ,

Metrical considerations
The particles -pa/ä(s), -s, and the optative mood in the data in this section are distinguished according to three categories: 1) tokens supporting the hypothesis of a moderative effect; 2) tokens that are ambivalent, potentially serving both moderation and for metrical requirements by filling the position of a syllable; 3) tokens inconsistent with the presented hypothesis. The criteria I have applied is based on an inspection of metrical factors of the verses with mentioned particles and optative verbs.

-pa/ä(s)
The data contains a total of 13 directives with particles -pa/ä(s), with apocope p' included. The forms -pa(s) and p' consist solely of verbs of requesting an agent to bring healing substance(s), and form eight tokens of the total 14. (20) Neitsy Maaria emonen Virgin Mary mother diminutive Tuopa hyytä hyppysisä Bring-pa frost with your fingers (SKVR I 4 , With six tokens, the form -pä exhibits more category-internal variation, where three tokens occur with verbs requesting an agent to perform movement, and another three with verbs of action. All of the tokens in the -pa/pä(s) category share their occurrence in verses addressing positive evaluated, supporting agents. The role of moderative effect of the particles of this category, however, cannot be established with certainty. Table 6   Table 6 Particle -pa/pä(s) and syllable count.

Particle
Tokens Minimum syllable count filled without the affix (number of lines) -pa(s) 4 1 -pä(s) 6 3 -p' 4 -reveals the problem: in four occurrences, line containing the directive would be metrically valid (at least eight syllables long) without an affix used. In ten tokens, omitting the affix would result in less than eight syllables in a verse. A question then arises: do the four tokens with "unnecessary" syllables support the moderative hypothesis? When the tokens are examined in context with the surrounding lines, only one (addressed to an unnamed agent) occurs in relation with other signs of possible moderation, namely a justification and an optative form (SKVR I 4 , 329). The other three not only lack other tokens of moderation, but furthermore contain the verb "(I) command" (käsken) in the same or surrounding lines. This kind of inconsistency does not support the moderative hypothesis.

Particle -s
The -s particle differs from the -pa/ä(s) in consisting of only one phoneme. It therefore lacks a similar metrical role with the syllable -pa/ä(s). Of the total 25 tokens in the data, 40% occur in directives aimed at supporting actors, like the Virgin Mary or the Honeybee. However, nearly the same amount, or 32% of the directives with the -s token are addressed to fire as a harming substance, which is told to leave the patient: (21) Nouses tuli taivoseh Fire, rise-s to heaven Auver pilvihi ajaite Smoke, flee to the clouds Sukuhusi syntyhysi To your kin, to your origin Heikko heimokuntahasi Weak one, to your people (SKVR I 4 , Relative to the examination of particles -pa/ä(s) in the context of signs of moderation in surrounding verses, nine of the tokens occur in relation to a justification or positive attributes. This indicates the possibility of moderation with a stronger basis than the case of mere particle use. When accounting for nearly the same number of tokens directed to fire as a negative agent, however, the moderative function does not seem that plausible.
As a result of this examination, I argue that the affix -s does not index pure moderation per se. Instead, according to the data, the affix seems to function in a similar sense as in standard language: as a directive to perform an immediate task (Hakulinen et al., 2004, x 383). However, it is possible that the effect of the affix can contextually indicate some degree of moderation, when addressing supporting agents.

The optative form
The total number of optative forms in the data is somewhat biased by its extensive use in one incantation text (SKVR I 4 , 285), in which 10 of the total 25 tokens occur. Still, all of the tokens in the data share their conformity to verse. An examination of the tokens reveals that they function at least as metrical optatives (Saarinen, 2018, 329-330), that is, they are means of achieving the minimum syllable count: (22) Käyös hyytä Pohjolasta Fetch opt frost from Pohjola Jäätä kylmästä kylästä Ice from the cold village (SKVR I 4 , The optative form occurs in directives aimed at fire, of which the status is ambivalent, in 32% of the tokens and in 68% when the performer addresses supporting agents. Notably, however, when fire is addressed as a negative agent, this is also done with optative forms, as well as being designated with positive attributes. This could indicate a connection between the optative form and moderation. However, due to the great relative number of optatives in one performer's incantation (SKVR I 4 ,285), it is hard to separate the possibility of moderation and aspects of an idiolect. I see the optative form in the data primarily as metrical filler, but as with the -s particle, it is with probability filling a secondarily moderative form.
The analysis of the particles shows that the initial hypothesis of particle moderation being similar to usage standard language is, for the most part, not supported by the data. This is the case especially with -pa/pä(s), where metrical reasons seem to dictate the usage. The -s particle and optative are more complex types: s is not used as a metrical filler, while the latter extends its function from metrical fillers to moderation in some contexts.

Discussion
This analysis provides a point of departure for further discussion about oral poetry's registers within an incantation genre.
The data I have presented shows that Origin of Fire incantations of Viena Karelia include in them varying register usages. These instances of differing registers, employed by the performer, are analyzable by a methodology that focuses on grammatical features and rhetorical devices.
My initial hypothesis was that the addresses of supporting and hostile agents could vary in usage of so-called moderative particles (-pa/pä(s), -s) and the optative form in directives, both of which could potentially be emblems of a moderative register. In addition, I examined justification of actions from the viewpoint of the hypothesis. The analysis resulted in two-fold results. The case of particles as moderative devices in directives was found to be mainly improbable. Statistical analysis does not support the idea that the two longer particles, -pa(s) and -pä(s), would unambiguously function as moderative devices. The -s particle and the optative form proved to be more complex in their roles. According to the available dataset, while they no doubt are utilized in a metrical sense, it seems at least plausible that they can also index contextual moderation. The role of justification as a potential part of the moderative register, on the other hand, is considerably more salient than the suffixes. This leads me to propose that, in the frame of this limited data set, justification may function as a moderative device, although this theory should be tested further against a larger dataset.
The results show that, in this data, the performers seem to conceive different register usages as social situations in ritual action. These situations are related to agents, which are addressed in the incantations, and the evaluations of these agents as either positive or negative. The findings about the registers of the Origin of Fire incantations carry significance that extend beyond just one ritual genre. As I have sought to demonstrate, reflexive registers can index not only social personae connected with them but also situations of interaction. This opens up chances of re-assessing the relationship of registers and genre in Kalevala-metric ritual poetry. One of the questions for future study is whether such registers also operate between genres of the poetic system.

Declaration of funding source
Doctoral Programme in History and Cultural Heritage, University of Helsinki has financed the research. The funding source has not influenced collection, analysis, or interpretation of data.

Declaration of competing interest
None.