Abstract
In studying ideophones in motion descriptions, this article focuses on Khalkha Mongolian (sov, agglutinative), a verb-framed language (Talmy, 2000) with occasional satellite-framed constructions and a rich inventory of manner verbs (cf. Matsumoto and Badema, 2025). Drawing upon a database of ideophones built within our project and supplemented by elicitation questionnaires, we (i) scrutinize the semantics of Mongolian ideophones in motion descriptions (partially using Ibarretxe-Antuñano’s (2019) Semantic Grid), and (ii) define the role of ideophonic constructions in forming core predicates or satellites in motion constructions (Talmy, 2000). We also examine the status of ideophones in relation to manner verbs, particularly if they are used in place of manner verbs (cf. Slobin’s (2006) discussion of manner salience) or as their secondary modifiers (cf. Schaefer, 2001). We have identified the prominent role of ideophones on the periphery of motion construction, whereas the contribution of Mongolian ideophones to manner salience remains disputable.
1 Introduction
This paper analyses ideophones in motion constructions in Khalkha Mongolian. It has the following goals: (i) to scrutinize the semantics of ideophones in motion descriptions (with the use of Ibarretxe-Antuñano’s (2019) Semantic Grid), (ii) to determine the role of ideophonic constructions (roots with suffixes and/or quotative verbs) in forming the core predicate or a satellite within a motion construction (in terms of Talmy, 2000), and (iii) to investigate whether ideophones have a specific role in relation to manner verbs, i.e. whether they are used instead of manner verbs (contributing to manner salience, cf. Slobin, 2006), or, conversely, whether they rather function as secondary modifiers of manner verbs (cf. Schaefer, 2001).
Mongolian belongs to the Mongolic language family. We focus on Khalkha, the official language of Mongolia, with ca. 3.5 million speakers. It has agglutinative morphology, and its default word order is sov. In terms of motion verbs, Mongolian is a verb-framed (Talmy, 2000) language with occasional satellite-framed constructions and a rich inventory of manner verbs (cf. Matsumoto and Badema, 2025). The Khalkha lexicon (as is true for all modern Mongolian languages and dialects)1 is usually described in terms of two main parts of speech – nominals and verbs. Verbal and nominal roots are distinguished, and four types of derivational suffixes are used to derive new words: deverbal nominalizers, deverbal verbalizers, denominal nominalizers, and denominal verbalizers.
This also applies to Mongolian ideophones, which are structurally complex and consist of ideophonic roots combined with derivational morphology. Ideophonic roots can take both nominalizers and verbalizers, and a special set of suffixes is used for them. Mongolian ideophonic roots can be (i) combined with suffixes and quotative verbs to form an analytic construction, “ideophones per se”, or (ii) they form ideophonic nominals using various nominalizers (not relevant for present study), or (iii) they derive ideophonic verbs by means of different verbalizing suffixes, which are specific to ideophones. Ideophonic verbs behave as regular verbs; however, they have semantic and pragmatic properties (iconicity, expressivity), which distinguish them from prosaic verbs, similarly to the “ideophones per se”. In motion descriptions, they appear on the periphery of the complex predicate.
Whereas in many other languages ideophones are often bare roots, or reduplicated bare roots respectively (as Childs (1994: 185) stated, “ideophones display very little morphology”), Mongolian ideophones are rather complex, corresponding to general morphosyntactic properties of a language. In this respect, we agree with Ameka (2001: 45) that “ideophones are an integral part of the languages in which they occur” and should be studied as such. The complex morphology of Mongolian ideophones has specific implications for the current analysis: Suffixes forming Mongolian ideophones have their own semantics. A vast majority of ideophones consist of a root and an obligatory suffix to form a stem. Surprisingly, this holds true even for reduplicated ideophones (cf. Ulman, 2024). Usually, it is not the ideophonic root that carries the motion semantics; rather, it is one of the suffixes or constructions, as will be discussed below in detail.
1.1 Previous Research on Mongolian and Mongolian Ideophones
Overall, Mongolian is relatively well studied, with the earliest descriptions dating back to the mid-19th century (Schmidt, 1831). Some early complex grammars of Written Mongol (e.g., Poppe, 1937) and comparative studies of Written Mongol and spoken Khalkha (Ramstedt, 1908) are still relevant. Of the spoken languages, Khalkha is by far the best described, both within comprehensive treatments of Mongolian (e.g., Janhunen, 2012), and in works focusing solely on Khalkha, such as descriptive grammars (e.g., Poppe, 1951; Luvsanvandan, 1968; Kullman and Tserenpil, 1996; Sechenbaatar, 2003), phonology (e.g., Svantesson et al., 2005), and book-length studies of individual aspects of grammar (e.g., Byambasan, 1970; Önörbayan, 2000; or Brosig, 2014). Today, various types of modern dictionaries include monolingual (e.g., Cevel, 2013; Bold, et al., 2008, etc.), bilingual (Russian, English, German, Japanese, Chinese, Korean, etc. – Mongolian and vice versa), and terminological (e.g., legal, geological, financial, etc.) dictionaries. Apart from these, there is a wide variety of dictionaries focused on different aspects of Mongolian linguistics, such as etymology (Sühbaatar, 1997; Tömörtogoo, 2018), abbreviations (Pürevjav, 2010), and dialects (e.g., Oirat, Dörvöd). It is worth noting that many of the frequently used dictionaries are accessible in online versions.
Given this generally good level of description of Mongolian, the ideophones in this language, in particular outside literature in the Mongolian language, are rather severely understudied. In Mongolian linguistics, ideophones have been approached from two main perspectives. The first, prevailing approach, based on the ideophone-specific derivational morphology and focusing on ideophonic verbs, began with Byambasan (1970), who defines ideophones and onomatopoeia as closed groups of roots that take specific verbalizers. Önörbayan (2000) gives an exhaustive list of ideophonic and onomatopoeic verbal roots. However, these works do not discuss ideophonic nominals and ideophones per se. Monolingual dictionaries since Cevel (1966) are in line with this approach, featuring the structure of a Mongolian ideophone as a root + derivational suffixes. Second, in some descriptive grammars, the ideophones per se (and onomatopoeia) have earlier been treated within either a broader group of particles or as adverbs (cf. Luvsanvandan, 1968). Sechenbaatar (2003) treats ideophones per se as an independent sub-class of words. Oberfalzerová (2009), for the first time, gives a unified description of ideophonic (and onomatopoeic) verbs, nominals, and ideophones/onomatopoeia per se.
However, no complex research on ideophones that would cover inter alia semantic aspects has been carried out so far, and our current project aims to fill this gap. In dictionaries, ideophones are sometimes presented without any explanation, merely cross-referenced to synonymous entries. As a rule, information on their usage is lacking, and examples are drawn from the belles- lettres or folklore, rather than from colloquial speech. This also concerns the dictionary of Mongolian ideophones by Narantuya (2021). Despite containing approximately 1,700 entries, the dictionary is based entirely on selectively drawn examples from belles-lettres and folklore. It lacks comprehensive explanations for these entries, let alone information on other properties of the ideophones, such as semantic nuances, expressivity, etc.
1.2 Theoretical Background
Regarding the theoretical basis of our study, we employ Talmy’s (2000) typology. Talmy (2000) focuses on the types of boundary-crossing motion based on the way languages code the path element of the event. Taking this typology as a starting point for our analysis, we look into the role Khalkha Mongolian ideophones play in motion events. Talmy (2000: 213–288) categorizes different Co-events as Precursion2 (the Co-event precedes the main Motion event but does not cause or assist its occurrence), Enablement (the Co-event precedes the main Motion event, and enables the occurrence of an event that causes the Motion but does not itself cause this Motion), Cause, Manner, Concomitance (the Co-event co-occurs with the main Motion event, but this activity in itself does not pertain to the concurrent Motion), and Subsequence (the Co-event takes place directly after the main Motion event, and is enabled by, is caused by, or is the purpose of that Motion event).
As has been formulated by Slobin (2006) and further extensively discussed in the literature (e.g. Toratani, 2012; Akita, 2017; Matsumoto and Akita, 2020), ideophones play an important, while complex role in expressing manner, cf. also Nuckolls, 2014. Manner salience (“the level of attention paid to manner in describing events”), according to Slobin, is something languages differ greatly in. Satellite-framed languages are, in general, supposed to express manner much more frequently than verb-framed languages (Slobin, 2006: 113). Slobin deals with the role of ideophones only briefly, presuming, however, that they play a role in compensating for the low manner salience of verb-framed languages. As Khalkha Mongolian is a predominantly verb-framed language, it is worthwhile to test this assumption. In this paper, we, consequently, focus on how Khalkha Mongolian ideophones combine with, or substitute for, manner verbs. The very notion of manner has been subject to discussions, as Slobin himself has admitted that neither he nor Talmy provides a clean definition of this term (Slobin, 2006: 62). Matsumoto and Akita (2020) provide a more fine-grained typology of manner expressions that we will refer to in this paper.
Hence, ideophones have been mostly studied for the ways they encode Manner. However, it has been recently shown by Ibarretxe-Antuñano (2019) that other components, such as Path or Figure, may form part of the ideophone semantics: In Basque, while 97% ideophones express Manner, 22.5% have Figure and 13% Path, as part of their meaning. This concerns the so-called “motion ideophones” (Ibarretxe-Antuñano, 2019: 151). In this paper, we discuss the role of Khalkha ideophones in expressing the individual components of Motion events. It has been stated that motion ideophones are common across languages (Ibarretxe-Antuñano, 2019: 137–8). We will show that the situation of Khalkha Mongolian ideophones is complex in this respect. In particular, Khalkha appears to have no ideophonic roots with default motional semantics, which complicates the delimitation and description of motion ideophones. As a result, we have used a slightly modified system based on the notions of Ibarretxe-Antuñano’s Semantic Grid to account for as many instances of motion descriptions in our data as possible. As has been mentioned, Ibarretxe-Antuñano’s proposal focuses only on motion ideophones, i.e., ideophones that primarily describe motion by themselves, while in our Khalkha Mongolian data, these ideophones are scarce. Therefore, this paper also deals with ideophones that are not necessarily motion ideophones but are often employed in motion descriptions and gain motion semantics in such an environment.
Akita (2017) has defined three parameters for the classification of manner expressions, which pertain to “three different facets of language: the lexicon, syntax-semantics interface, and syntax-pragmatics interface” respectively: (i) Grain size, (ii) Event inherence, and (iii) Expressiveness. He proposes the following three hypotheses concerning the three parameters: (i) Coarse-grained manners (e.g., running, flying) are more likely to be lexicalized in verbs than fine-grained ones (e.g., bustling, swooping). (ii) Manners inherent to motion events (e.g., running) are more likely to be encoded in tight motion constructions than those that are only concomitant with them (e.g., dancing, buzzing). (iii) Intensified manners (e.g., running at a surprisingly fast rate) are less likely to be encoded in verbs than plain manners (e.g., running at a normal rate). In our analysis, we will also look at these parameters.
To date, Mongolian motion ideophones and ideophones in motion descriptions have not been analysed from the perspective of Path and Manner coding. Oberfalzerová (2009, 2010, 2011) describes examples of ideophones in motion constructions and the semantics of the individual verbalizers, analysing them from a pragmatic and ethnolinguistic perspective. However, her account does not focus on coding elements of Motion. There is Matsumoto and Badema’s general analysis (2025) of motion descriptions in Khorchin Mongolian, discussing Path coding in this Mongolian dialect, and showing, among others, that Khorchin has higher manner salience than is typical for other verb-framed languages. Khalkha Mongolian behaves in a similar manner.
The structure of this paper is as follows: In Section 2, a short overview of Mongolian ideophones is provided, including their morphosyntactic features and semantics. In Section 3, the data and the methodology used in this research are discussed. The main part of the paper, Section 4, then consists of the results of our analysis, first the detailed discussion of semantics of ideophonic roots, suffixes and the distribution of ideophones across semantic domains (4.1), then discussion of their syntactic properties as seen in the data (4.2), followed by the role of Mongolian ideophones in manner salience (4.3), and overall discussion notes (4.4). Section 5 contains our conclusions and implications of our findings.
2 Morphosyntax and Semantics of Khalkha Mongolian Ideophones in Motion Descriptions
In Khalkha Mongolian normative and descriptive linguistics, ideophones (dürsleh üg3 ‘depictive words’) and onomatopoeia (avia duuraih üg ‘sound- imitating words’) are usually treated as two subdivisions of a separate closed word class broadly defined by employment of iconic features and by specific morphology. The onomatopoeias are out of the scope of the present study.
2.1 Morphosyntax
The word class consists of several hundred roots, ca. 500 ideophonic and 200 onomatopoeic according to Önörbayan (1988, 2000). In our database, we have registered 889 ideophonic and 201 onomatopoeic roots so far, from which nominals and verbs are formed by derivational suffixes and their combinations. Nominals (i.e., pos in an attributive, adverbial, and nominal position) are mostly formed by the suffixes -gAr,4 -uur, and -gAi/-hAi. Verbs are formed with the suffixes -Ai- (static) ‘to have or acquire the quality described by the root’, e.g. band-ai- to be/become plump with broad bottom’, -lzA- (iterative) ‘repeated movement’, e.g. band-a-lz- ‘to move repeatedly of a plump shape’, and -gAnA- (iterative) repeated regular movement’, e.g. band-gana- ‘to perform regular movement of a plump shape’. Ideophonic verb stems, formed in this way, take regular suffixes of verbal inflection.
In addition to verbs, ideophones per se (adverbially used nominals) occur in motion constructions (quotative constructions in Akita, 2017). The suffixes forming these nominals are -gA, -s, -r, -n, and -rs. The latter three are rather rare, occurring in less than 1% in our corpus, therefore, they are not included in the analysed sample. The ideophones with the suffix -gA, mostly reduplicated, are either followed by a do/say5 verb or used independently, cf. the elicited examples in (1). The suffix -s most often expresses a one-time movement and typically occurs with the ‘do’ verb, as in the elicited examples in (2).






The following is the list of the most frequent suffixes and constructions occurring in motion descriptions:



It has been indicated by speakers, as well as noted in dictionaries, that the iterative (regular) verb -gAnA- (i) is synonymous with the analytic constructions (ix-xi), and the iterative (irregular) verb -lzA- (ii) is synonymous with the constructions (vi-viii). For this reason, we treat ideophonic verbs in this paper as “true” ideophones similar to the ideophones per se.
The resulting forms (root + suffix or analytic construction), as such, do not express motional meanings. As stated by all our consultants, these forms overwhelmingly refer to movement that does not involve Path. In motion descriptions, they are typically followed by (a) motion verb(s) to achieve a motional meaning.
Additionally, several sound-imitative ideophones (or onomatopoeia) occur in motion constructions. Sound-imitating roots have a slightly different morphology: Unlike ideophones “proper,” they are used as bare roots (e.g., shur ‘rustling sound’ > shur hii-/ge- ‘produce a rustling sound’. Verbs are mainly derived by the suffix -gInA- expressing continuous sound, e.g. sün ‘whoosh’ sün-gene- ‘to whoosh continually’.
An interesting device that occurs in motion descriptions is the use of causative. It is optional and has been explained by our consultants as “the agent of motion showing (=causing being seen) a shape in motion” by moving him/herself, e.g., (3).



Typically, the causative is also applied when the agent is using a means of transport (horse, car, etc.), as in (4).



The use of causative is a device used in motion descriptions in general, not only with ideophones.
The stative suffix -Ai-, when occurring in motion descriptions, is frequently attached to ideophones with shape semantics, usually combined with a causative suffix (example (3)). However, it is also typically used with one of the roots referring to motion, -jir- (jirii-) ‘to perform rapid and straight motion’).
As an anonymous reviewer has suggested, an analogical pattern of formation, i.e., an ideophonic root + suffixes encoding semantic parameters similar to Mongolian, in addition to expressive means such as reduplication and vowel alternation, exists in Rgyalrongic. In Japhug (Jacques, 2013: 258–262), there are nine morphological patterns (including the bare root), six of which are dynamic and may involve motion. The semantics of the morphological forms is close to those of Mongolian. Interestingly, however, in Japhug the bare root carries dynamic (semelfactive) meaning, which is not true for most Mongolian ideophonic roots. Unlike Mongolian, the Japhug ideophones do not seem to form verbs.
A rather similar system also exists in Nanaic (Tungusic). Petrova (1948: 529) distinguishes two types of ideophones in Nanai and Ulcha: (i) monosyllabic roots which employ only expressive morphology (reduplication), and (ii) disyllabic roots, employing a set of ideophone-specific suffixes deriving ideophones per se, which can further be reduplicated. Petrova lists over 10 different suffixes with both stative and dynamic meanings, at least two of them implying motion. Similarly to Rgyalrongic, these forms are mostly used adverbially but are not verbs themselves, unlike in Mongolian.
On the other hand, Sakha (Turkic, Siberian) ideophones, similarly to Mongolian, employ verbalizers, roughly half of which are ideophone-specific, expressing meanings similar to Mongolian, including those used with motional semantics (Kharitonov, 1982). The similarity of the Sakha system to Mongolian has been ascribed to contact (e.g., Pakendorf, 2015).
The three cases listed above show that rich ideophone-specific morphology is not an exception, although ideophonic systems that prefer expressive means such as reduplication seem to be more common in the wider area of Northeast Asia (Ulman, 2024). Interestingly, in all the above-mentioned cases, motional meanings seem to be carried by non-sound-symbolic morphological means, rather than by the ideophonic roots.
2.2 Semantics
The semantics of ideophones in motion descriptions consists of two parts corresponding to their morphology: (i) the semantics of the ideophonic root (shape of a figure, e.g., cünd- ‘a compact, moderate-sized round shape’, sound, e.g. sün- ‘whoosh’, a visual effect, e.g. tor- ‘a bulging object in distance’, light, e.g. gyal- ‘lightning, shine’, character of surface, e.g. göl- ‘smooth and shiny’ or derj- ‘hard and uneven’, (ii) motional semantics of a suffix and/or a an analytic ideophonic construction comprising a ‘do/say’ verb (ideo-gAnA- ‘regular repeated movement’, ideo-lzA- ‘irregular repeated movement’, ideo-s + ‘do/say’ ‘one-time movement’, ideo-gA + ideo-gA + ‘do/say’ ‘repeated movement’, etc.).
Due to their transparency, Mongolian ideophones are usually described in terms of the semantics of their roots. The roots of ideophones cover all five main sensory domains (e.g. Winter et al.’s (2017) classification: visual, auditory, tactile, olfactory, gustatory). The distribution across domains is, however, uneven: More than 60% of all ideophonic roots belong (i) to the visual domain, which can be further divided into (a) roots referring to shape, such as bönd- ‘round shape’, band- ‘plump shape with broad bottom’, danh- ‘vertical shape with top broader than bottom’, goz- ‘prolonged erect shape’, etc. and (b) roots referring to visibility, effects of light, and arrays of objects such as gyal- ‘glittering, lightning’, myar- ‘shimmering, gleam’, tor- ‘prominence in distance’, jir- ‘tiny objects ordered in a straight line,’ etc. Group (a) also codes hardness-softness, and flexibility, of the shapes, however, within the visual domain (cf. Oberfalzerová, 2009). The auditory domain (ii) follows with some 20–25% of the whole inventory of roots. The tactile domain (iii) may contain several tens of roots expressing properties of surface, such as barz- ‘rough’ or göl- ‘smooth and shiny’. However, it has been noted by our consultants that even in this category, it is often the visual perception (‘rough looking surface’) that is primarily evoked by the ideophone. Finally, the olfactory (iv) and gustatory (v) domains are represented by probably less than a total of 10 items, such as han- ‘intensive smell’ and ülb- ‘tasteless’.
Hence, Khalkha Mongolian does not have ideophones primarily pertaining to other sensory domains (the vestibular (balance) and the interoception (bodily feelings) domain). Unlike some other languages with rich ideophone inventories, such as Japanese (Hasada, 2001) or Basque (Ibarretxe-Antuñano, 2017), Mongolian does not have primary ‘psychomimes’ – ideophones describing emotions and psychic states, and – importantly – does not have roots which primarily refer to motion. Mongolian, therefore, broadly follows the implicational hierarchy proposed for semantic domains of ideophones (as in Dingemanse, 2012).
That said, Khalkha ideophones are eventually used in the abovementioned domains as a result of the semantics of their morphosyntactic formants, semantic shifts, and lexicalization.
3 Data and Methodology
To achieve the three aforementioned goals, the following topics are scrutinized in this study:
- (i)we analyse semantics of ideophonic roots using Ibarretxe-Antuñano’s (2019) Semantic Grid, i.e. we make use of three-level hierarchy of semantic components defined in Ibarretxe-Antuñano (2019: 149): seven first-level components (figure, ground, path, motion, manner, cause, event extension) subdivided into 16 second-level (quanta, animacy, physical characteristics; type, physical characteristics; contour, direction, deixis; movement, stationary; motor pattern, energy, rate, force, regularity, step, leg, instrument, sound, posture, figure-induced behaviour, casualness; causer, reason; phase, aspect) and 121 third-level subcomponents;
- (ii)we survey the syntactic properties, i.e., the role of ideophonic constructions (roots with suffixes and/or quotative verbs) in forming the core predicate or a satellite within a motion construction (in terms of Talmy, 2000);
- (iii)we test whether ideophones have a specific role in relation to manner verbs (manner salience or secondary modification of manner); based on our preliminary investigation we do not presuppose a specific role, we expect that ideophonic roots can form constructions combined with both path and manner verbs, and they can form as a core a motion construction on its own; alternatively, two assumptions suggesting functional tendencies stated in the literature are applicable for Mongolian: (i) that ideophones in verb-framed languages are used instead of manner verbs (Slobin, 2006), i.e. they occur predominantly in combination with path verbs, or (ii) that ideophones rather function as secondary modification of manner (“tier-two manner senses”, Schaefer, 2001), i.e. they combine chiefly with manner verbs.
A database of ideophones with analysed examples is being created within our project, containing multiple examples from the main registers in which ideophones are used. Data sources include different spoken and written registers (colloquial language, journal and newspaper articles, literature, folklore). We mainly rely on fieldwork data, the greater part of which was collected by A. Oberfalzerová and J. Luvsandorj between 2000–2010 (producing ideophones used in spontaneous speech) and again by a multi-sited fieldwork both in rural and in urban areas in 2022–24 during which 30+ speakers of different background and age were interviewed for a set of ca. 100 ideophones. The latter part of the fieldwork was conducted in the fashion of half-structured interviews, partly with groups of speakers, and it has yielded, besides elicited examples, a number of spontaneously used examples. Other data sources involve literature (fiction), journal and newspaper articles, and folklore collections. Finally, we have supplemented our data with examples collected from social media and comments to internet articles.
For the purpose of this study, we have used, to the greatest possible extent, data from natural spoken discourse (fieldwork data, supplemented with examples from social media which can be identified as colloquial through morphological and orthographical features). Additionally, we used several examples from the literature after having verified their usability in colloquial language with our consultants. We made a qualitative analysis of the whole data set, selected diverse types of ideophones that co-occur with the verbs of motion or form the core predicate of a clause describing motion, and analysed their properties.
For our analysis, we considered ideophonic roots that describe motion within ideophonic stems (i.e., in a combination with derivational suffixes), and ideophonic roots that do not express motion on their own but appear readily in complex constructions with verbs of motion.
For goal (i), we have devised, drawing on our database, two questionnaires6 with two different sets of ideophones and consulted them with native speakers for different purposes:
- (A)Questionnaire 1 (semantics): a set of 100 ideophonic forms (deriving from 100 different ideophonic roots) which have occurred more than once in motion descriptions in our data. The base of this sample was formed by ca. 50 roots occurring in motion descriptions (either in natural discourse, or naturally (i.e., without specific instruction) given during elicitation focused on ideophone semantics) in our fieldwork data (Oberfalzerová, 2009, 2010, 2011, and recent fieldwork data collected between 2022–2024). Aiming at a sample of 100 ideophonic roots, we searched our database for ideophones (regardless of their source) for which motional use was indicated in translation or given in examples, and we verified the motional use via consultations with native speakers. The questionnaire was composed of (i) 5 roots (the total number of this type in our data by then) for which preliminary consultations suggested prevalence of motional use, (ii) 10 roots for which, based on data and consultations, we supposed denial of motional use, and (iii) 85 roots without presupposition of results. Reflecting the fact that ideophonic verbs are more frequent in contemporary spoken discourse, the questionnaire set consisted of 65 ideophonic verbs and 35 ideophones per se. Afterwards, this set was consulted with 6 native speakers to establish which ideophones are most likely to occur in motion descriptions. The speakers were asked to give the first example sentence that came to their mind, and (optionally) comment on the semantics of the ideophones. To minimize the influence on the respondents’ decisions, the ideophones were given in their dictionary forms (future participle for ideophonic verbs and bare forms of ideophones per se).
- (B)Questionnaire 2 (morphosyntax): A selection of 20 ideophonic roots, each in 5 main forms, including variants (11 forms for each ideophonic root), was consulted with 3 native speakers to gauge the ability of the individual roots to combine with derivational morphology. The selection included 10 ideophonic roots with supposed motional semantics (see below), and 10 ideophones with non-motional semantics which, nevertheless, have been attested in motion descriptions.
For goals (ii) and (iii), we used (C), as the main data:
- (C)Motion description sample: A set of 100 examples of ideophones in motion descriptions containing 70 ideophonic roots. The examples were selected from (a) fieldwork data, (b) internet resources (e.g., comments to articles). The sample included 70 ideophonic verbs, 25 ideophones per se in constructions with iterative meaning, and 5 ideophones per se in the construction expressing one-time movement – a ratio which corresponds to the distribution of forms of ideophones in motion constructions across our database. This sample has been analysed in terms of linear position and syntactic properties of the ideophones, their role in coding Manner, and the type of co-event they express. Further, we consider this sample as reflecting the approximate relative frequency of ideophones not only in terms of form, but also in terms of semantics.
4 Analysis
4.1 Semantics of Ideophones in Motion Descriptions
In this section, we deal with the semantics of Mongolian ideophones: the co-event they represent, the semantic domains they cover, the semantics of the root and the suffixes they contain, and finally, we analyse in detail the semantics of the set of ideophones using the Semantic Grid.
As a starting point, we analysed sample (C) – a set of 100 motion descriptions – in terms of the relationship between the co-event and the main motion event (based on Talmy, 2000). We concluded that, of Talmy’s eight types of co-events, 83 of the analysed examples are best described as Manner, while 13 instances can alternatively be considered Precursion, as in example (5). Coding Precursion in Mongolian appears to differentiate Mongolian from other verb-framed languages such as Japanese, in which the possible co-events of motion are limited to Concomitance, Manner, and Concurrent result (Toratani, 2012: 115). Furthermore, several cases (at least five, including example (6), and potentially many more depending on interpretation) could alternatively be classified as instances of Concomitance.






One result of our analysis of ideophones in motion constructions was a surprising feature of the Mongolian ideophone inventory in comparison to other ideophone-rich languages such as Japanese or Basque. In Japanese, the ideophone sutasuta ‘walking briskly or hurriedly’
(Toratani, 2012: 122), and in Basque, the ideophone bristi-brasta ‘walk hurriedly’ (Ibarretxe-Antuñano, 2019: 145) both express a motion event (walking). In Khalkha Mongolian, a similar meaning may be expressed with the ideophonic verb toro-lzo- ‘to walk briskly’. However, this verb is, derived from the root tor- ‘to be seen from afar’ and the main meaning of the form in question is ‘to be seen from afar, moving”. The motional meaning is figurative and arises only in certain context. Given the fact that motion ideophones are “ubiquitous” in languages (Ibarretxe-Antuñano, 2019), and that Khalkha Mongolian has several hundreds of ideophonic roots and several thousands of individual forms, it was unexpected to find out that there are no roots with motional semantics, and very few stems that describe motion independently of context.
As mentioned above, the semantics of Mongolian ideophones is compositional, the meaning of an ideophone deriving from the meaning of its root and its morphosyntactic formants. Consequently, ideophones used in motion constructions can be dissected and the individual components described in detail.
Motional meanings are always a result of a semantic shift, and/or lexicalization, as is illustrated by the ideophonic verb god-gono- ‘skip, move without effort’. This verb derives from the root god-, referring to a small, upwards- sticking object. Adding the repetitive suffix -gAnA- results in the meaning ‘regular movement of a small, upward-sticking object’, and as such is applied, for example, to movements of a goat’s tail. The ideophone is further used, in a figurative meaning, to describe the behaviour of lively, fidgety children, especially girls. Finally, it has been lexicalized from the latter usage to refer to light, effortless movement or motion, as in (7).



Hence, Khalkha ideophones that occur in motion descriptions are always derived from stems with non-motional semantics, as the Motion descriptions (C) sample clearly shows. Of the 100 examples of motion descriptions, 90 instances are derived from roots belonging to the visual domain, overwhelming majority of such roots refer to shapes (cünd- ‘round and compact’, cülh- ‘round and jelly-like’, togd- ‘short and compact’, gulbi- ‘long and elastic’, etc), or (in just two instances) to properties of surface (göl- ‘smooth’ and derj- ‘wrinkled’) and, in another two cases, visual effects (tor- ‘a small bulging object visible in the distance’). Ten of the roots refer to sounds (e.g., sün- ‘whoosh’). The ideophones in the sample take either verbalizing suffixes (70 instances, resulting in the meaning of repeated movement, eventually state), or, as ideophones per se, appear in constructions with the verbs hii-‘do’ or ge- ‘say’ (22 instances resulting in repeated or one-time movement). Finally, in eight cases, ideophones per se directly modify a verb.
Most of the resulting stems describe movement (without motion) and, to achieve the motional meaning, need to be combined with manner or path verbs. During interviews with consultants stems have been unambiguously explained as “shapes/visual effects that move repeatedly on the spot in a manner that is typical (and possible) for such shape”, meaning that, for example, the ideophonic verb bond-gono- ‘round shape + repeated movement’ can, depending on context, refer to walking, running, jumping or rolling of the Figure, but will not refer to swaying, which is a type of movement associated with long shapes.
During interviews conducted within fieldwork 2022–2024, however, a few ideophone forms were repeatedly described as referring to the speed of motion. Further, certain forms with repetitive formants turned out to occur in motion descriptions more regularly than others. To gain a complex view on what types of ideophones exist in spoken Khalkha Mongolian and how they are distributed, we analysed the (C) sample of 100 authentic motion descriptions containing ideophones in 100 distinct ideophonic senses (derived from 75 different ideophonic roots).
This sample has pointed in the direction that, in terms of motional semantics, instead of forming two clear-cut groups, four types of ideophones with varying contribution to the motion description can be distinguished, and these groups may be arranged in a “motion lexicalization scale”, i.e. integration of motional meaning into the ideophone stem: starting with (4.1.1) ideophonic stems (shape + verbalizers/quotative constructions) which have no motional meanings as such and assume motional semantics set only in a sentence with motion verbs and/or context), over manner modifications (4.1.2), path modifications (trajectory) (4.1.3) and ending with group (4.1.4) expressing inherently motional meanings (speed).
In order to distinguish these from the majority of non-motional ideophones used in motion descriptions, we used (A) Questionnaire 1. We consulted the selection of 100 ideophones which had been attested in motion descriptions with six speakers (four male and two female, in the age range between 19 and 61 years, of whom three grew up in the city of Ulaanbaatar and three grew up in the countryside). The speakers were primarily asked to give the first example with the ideophone in question that came to their mind. Additionally, the consultants were given space to comment on the semantics of the ideophone – in the example they gave, as well as in general.
The consultants, generally, agreed with one another on the distinctions among these groups. They were expressed in several ways: (i) explicit statements regarding semantics as e.g., ‘(permanent) shape, which performs repeated movement’, (ii) description by synonyms (manner verbs, e.g., sway, limp, shake), (iii) gestures.
Analysing results of (A) Questionnaire 1, we counted examples with clear motional semantics, excluding all potentially ambiguous cases, and based on the survey, we selected 16 ideophones for which examples involving motion were given predominantly (in at least four responses of six). An example clause was considered a motion description when it (i) contained a clear expression of Path (e.g. ‘dashed to school’ or (ii) when the ideophone modified a motion verb, or a sequence of motion verbs (Path, Manner and/or general motion verb ‘to go’).7 Ambiguous cases mostly contained the verb of general motion hödlö- ‘to move’, which, in principle, can refer to both movement and motion. As a result, we thoroughly analysed the 16 selected ideophones using the Semantic Grid. For the purpose of analysis, we have conducted additional interviews with native speakers, which further clarified certain issues.
4.1.1 Ideophones Expressing a Shape in Motion (26 of 100; 2 of 16)
A group of ‘shape’ ideophones encompasses 26 of 100 instances of the (C) sample and 2 of 16 ideophones with clear motional semantics, analysed in (I.1)–(I.2) below. The ideophonic roots refer to shapes. Both verbal derivations of these roots and analytic constructions with ideophones per se express movement, most often in the form of a converb that modifies a motion, path or manner verb, cf. (8).



The root band- refers to a rounded shape with a broad bottom, in most cases referring to fat people or animals, or their body parts (e.g., the fatty tail of the Mongolian sheep). When used in motion descriptions, it may be combined with virtually any of the suffixes and constructions of (i)–(xiv), which express either repeated movement or one-time movement. The resulting form evokes a visual perception of a movement of such shape, i.e., either of a fat body or a body part moving. These forms do not imply any motion performed by the person or animal possessing the described property. Only when the resulting verbalized form combines with a general motion verb (to go) or a manner verb, it may be interpreted as referring to motion. When combined with a path verb, it refers to Motion. Our consultants have repeatedly interpreted it as movement which is typical, and possible, for such a shape (e.g., can be interpreted as walking, running, shaking, etc, according to the context, but would not be interpreted as swaying or wriggling, because such movement is only possible for long shapes).
This semantic type can be interpreted as assigning properties. More specifically, the shape described represents a permanent property of the Figure, therefore, we classify it as Figure according to the Semantic Grid. ‘Shape’ ideophones typically occur independently of motion events. In responses to (A) Questionnaire 1, speakers, when commenting on this type, invariably described it as a (permanent) shape moving repeatedly, which may happen in the course of, e.g., walking or running, but also otherwise, sometimes stating that the occurrence of these types of roots in motion descriptions is purely accidental. Descriptions of this type have been accompanied by gestures forming a shape (round, long, pointy, etc.), sometimes illustrating the meaning by pointing at (or grabbing) body parts (belly, ears, nose, etc.).



In (B) Questionnaire 2, this group of ideophones showed the greatest morphosyntactic variability, combining freely with all 5 basic types of suffixes and constructions. Hence, this type can be considered to have the lowest degree of motion semantics and the loosest connection to the Path element. In terms of Talmy’s co-event classification, the most frequent relationship is Manner (alternatively Concomitance), but it is also only this type that may be classified as Precursion (analysed in the sense that the shape in question was formed before the motion event and independently from it). As for Ibarretxe-Antuñano’s framework, this type is the only one that typically involves Figure. Following Akita’s (2017) parameters, this type mainly expresses a fine-grained manner and, accordingly, is the least likely to be lexicalized in verbs (i.e., this type occurs in the “ideophone per se” constructions relatively more frequently than the rest.)
4.1.2 Ideophones Expressing a Specific Modification of Manner (56 of 100; 6 of 16)
‘Manner modification’ ideophones form the absolute majority, 56 ideophones, in the (C) sample. Six ideophones predominantly exemplified by our consultants are analysed in (ii.1)–(ii.6). Ideophonic manner-based roots are combined with a suffix expressing movement. Their modification mostly results from the perception of posture and its role within motion (e.g., haz- ‘leaning to one side’, haz-gana- ‘to limp’), but it may also be a general property of motion (düüh- ‘energetic, striving’, düühe-lz- ‘to walk/run in an energetic manner’). For three roots, we identified a sound-symbolic etymology, figuratively shifted to refer, e.g., to energetic walk in (ii.6).



The root sair- (sari-) in example (9) refers to something long and crooked and is typically used for legs. When combined with the suffix -gAnA-, which often implies walking, it refers to waddling or wobbling while walking. This movement likewise does not involve motion but is less static than the preceding type. It refers to motion when accompanying a manner verb and to boundary-crossing motion when accompanying a path verb.
We analyse these ideophones as expressing Manner with specific modificational subcomponents within the Semantic Grid. The most frequent modification is Manner: Posture (with dynamic characteristics, i.e., rather “positioning”).



In terms of relative frequency, this type occurs more typically in motion constructions than the preceding one. When commenting on examples of this group, the speakers used manner verbs to describe them (‘sway’, ‘bend’, ‘bow’, ‘shake’, etc.) and described them as repeatedly changing shapes. Descriptions were typically accompanied by repeated, multiplied gestures of hands, feet, or repeated movements of body parts (e.g., repeated bowing or multiple movements intended to express high or low energy). Further, ideophones typically describing walking were mimicked using two fingers of a hand, in the cases of ideophones derived from sound-imitating roots (tev tev ‘energetic walk’, turjig tarjig ‘sound of horse hooves in gallop’ accompanied by drumming with fingers).
In (B) Questionnaire 2, this type shows slightly lesser variability (average 3–4 of 5 basic morphosyntactic forms). Concerning Talmy’s Co-event classification, the typical relationship is Manner or Concomitance.
4.1.3 Ideophones Expressing a Trajectory of Motion (5 of 100; 3 of 16)
‘Trajectory ideophones’ are represented five times in the sample; however, three stems were selected in (A) Questionnaire 1, cf. analysis in (iii.1)–(iii.3). Ideophonic stems express a modification of path (e.g., zigzagging).



The root haiv- in example (10) refers to a shape with its base smaller than its upper end, and its upper end being irregular (e.g., a vessel with an irregular rim). When used in movement constructions, it refers to a swaying movement (without inherent reference to motion). However, in motional contexts it can also refer to path. Unlike the previous case, the movement is performed by the whole agent. It is frequently (even perhaps typically) used in motion descriptions, joined by motion verbs and describing both non-boundary crossing motion and boundary crossing motion depending on the type of the verb.
We consider these ideophones to express secondary Path: Contour.



The ‘trajectory’ type has been rather rare in all our samples; in addition, all these roots are polysemous between ‘manner modification’ and ‘trajectory’. We distinguish the two senses, since the speakers illustrated trajectorial occurrences by gestures modelling (winding or zigzag) contours with palms or fingers, clearly indicating that the ideophones of these examples are closely connected to the Path.
In terms of Talmy’s classification, these are instances of the Manner co-event, while in the Semantic Grid, they pertain to Path. In terms of Akita’s (2017) grain parameter, the type of manner expressed is relatively coarse-grained. All of the occurrences in our corpus are verbs, suggesting a greater degree of lexicalization into verbs.
4.1.4 Ideophones Expressing the Speed of Motion (12 of 100; 5 of 16)
‘Speed’ ideophones are of the greatest interest for this paper, therefore we list examples of all the five selected ideophonic stems as they occurred in the sample (within 12 instances), together with their analyses (iv.1)–(iv.5). Instances of this type (one with the lowest frequency in our sample) have been identified by consultants as having clear motional semantics.






The ideophonic verb jir-ii- (root jir- ‘straight line of tiny objects’+ the stative suffix -Ai-), is being used in the figurative meaning ‘rapid, almost straightforward motion’ and lexicalized in the meaning of high-speed motion on the ground. In example (11), jir-ii-is the main verb and is modified by a path verb uruuda- (‘move downstream’) in its converb form. This construction is specific in that it is actually satellite-framed.






The root ood- in example (12) refers to something short. The verb ood-gono- (formed with the verbalizer gAnA ‘regular repeated movement’), primarily refers to a regular movement of something inappropriately short. When the agent is a cattle or horse, it refers to a specific situation when the animal, bothered by flies, raises its tail upwards and gallops. However, currently it is used more generally to describe galloping cattle.






The root sün (‘whoosh’) in example (13) is sound-symbolic. The suffix -gene- is a form of -gInA- (continuous sound). The form sün-gene- refers primarily to the howling wind or sound of an object flying in high speed, such as a flying arrow, and, in a derived meaning, to the motion itself. It is also used to describe high speed of motion, most often of cars and people.






The root shur in example (14) is an onomatopoeic root referring to a rustling sound. The construction with “do” is frequently used to evoke the image of a quick, rushing movement, which, however, may refer to a motion along a path. In this example, the sentence with shur hii- as a main verb is immediately followed by a sentence specifying the final point of the motion event.






The form uha-s hii- in example (15) is a lexicalized construction referring to a sudden movement up and forward, which is often a commencement of motion. Figuratively, it can also be used for the motion that follows the move-off, describing the motion as fast.
The ‘speed’ type is the only one interpreted in terms of motion by the consultants, in particular in terms of speed (‘fast movement, very fast movement, fast walk’, etc). Description was invariably accompanied by gestures of palms moving fast along a straight horizontal line. In the Semantic Grid, this type falls clearly within the Manner: Rate category. In terms of Akita’s (2017) parameters, this type expresses a rather coarse-grained manner, has the lowest degree of expressivity, and is frequently lexicalized in verbs. Questionnaire 2 shows that there is little morphological variation – at most 2 of the 5 basic forms were indicated by the speakers.
4.1.5 Conclusions for Semantic Analysis
Combining this data with the results of (B) Questionnaire 2 suggests that there are differences among the groups defined above in terms of their ability to take on different suffixes: Ideophones of group 4.1.1 combine rather freely with all the basic 5 types of suffixes and constructions, including variants and combinations. The applicability decreases with the increasing degree of motion semantics, the group 4.1.4 occurring with 1–2 types corresponding to the particular meaning of the ideophonic root. In terms of Talmy’s distinction between verb-framing and satellite-framing constructions, groups 4.1.1–4.1.3 are typically satellites (converbs modifying the main path/manner verb or another modifying manner converb within a construction). Group 4.1.4 frequently functions as the predicate core, i.e., ideophonic manner main verb. The four groups also form a scale in several other respects. In terms of Akita’s (2017) parameters of manner, Group 4.1.1 comprises ideophones that express the finest semantic distinctions, have the lowest degree of integration into motion descriptions, and display the greatest degree of expressivity. In accordance with Akita’s hypothesis, they virtually never form sentence heads. Conversely, the ideophones of the ‘speed’ group 4.1.4, which are the most likely to form core predicates, express a rather general manner of motion, are the most tightly integrated into motion events, and are the least expressive, thus supporting Akita’s hypotheses.
4.2 Syntactic Properties of Ideophonic Constructions in Motion Descriptions
In terms of Talmy’s (2000: 21–146) typology of lexicalization patterns, Mongolian ideophones roughly correspond to a Manner semantic element in motion constructions. As in some other verb-framed OV languages, such as Japanese or Korean (Matsumoto and Kawachi, 2020: 12), the order of members in motion constructions is usually Manner-Path-Deixis. Talmy (2000) considers Deixis a component of Path, and deictic verbs as a subtype of Path verbs in example (16). Matsumoto and Kawachi (2020: 12) note that in some languages, such as Korean, Japanese, or Thai, deictic verbs occupy a special slot in a verb complex and combine with other path verbs, and they treat them as a special group on the level of manner and path verbs.



When, in descriptions of non-translational motion, the verb of general motion yav- ‘go’ is part of the predicate, it forms the core and can be preceded not only by Manner verbs, but also by Deictic (not Path) verbs, e.g., example (17).



The verb yav- in such cases can be analysed as an auxiliary expressing durative aspect, however, it retains part of its motional semantics.
The syntactic analysis showed that ideophones occupy both (i) a peripheral and (ii) a core position in the clause.
- (i)Ideophones appear typically as satellites (87/100 instances), sharing the position with nominal and adverbial modifiers, i.e., on the left periphery of the motion construction. The usual order of members in motion descriptions is thus the following:
ideophone / satellite – manner verb – path verb – deictic verb – general motion verb
In such cases, the core slot may be occupied by manner/path/deictic verbs (which may be combined with the verb of general motion). Satellite modifiers, as well as other adverbial modifiers, are placed either preceding the ideophone (in 24 of 87 cases), or (in 10 instances) immediately following the ideophone, see example (18).



Two or more ideophones may combine in a clause: with the same root, such as example (19), with distinct stacked roots, as in example (20). Semantics of such constructions will be analysed in the following section.






- (ii)Ideophones may also form the core predicate (13 of 100 instances). In this case (in our sample) it is never modified by any verb, and usually has a satellite modifier which precedes it, as in example (21).



Hence, the typical position of Khalkha Mongolian ideophones is on the periphery of the motion construction, but they can also form the core predicate. In terms of manner coding, Khalkha ideophones can modify any of the types of motion verbs – a general motion verb, a deictic verb, a path verb, and a manner verb. Among the latter, ideophones modify not only Talmy’s “general” and Matsumoto and Akita’s “default manner” verbs, but also Talmy’s “specific” (Matsumoto and Akita’s “conventional specific”) manner verbs. The differences in meaning among syntactic constructions will be studied in the following section.
4.3 Semantic Properties of Constructions with Ideophones in Relation to Manner Salience
This section focuses on the semantic properties of ideophonic constructions emerging from distinct relations and roles within syntactic constructions. To test the assumption concerning manner salience and the role of ideophones therein, we focus on the distinct ways of manner coding in the clause: the role and variety of manner verbs (core predicate or satellites) and non-verbal manner satellites (nominal modifiers).
Following the division of Akita and Matsumoto (2020) of manner verbs into four categories, Khalkha Mongolian has not only a substantial inventory of “default specific” manner verbs (2 verbs for jumping, at least four verbs for swimming and floating, etc.), but also a relatively rich inventory of conventional specific manner verbs, such as get- and geldre- for slow and cautious walk, or davhi-’ gallop’, shogshi- ‘trot’, hatir- ‘fast trot’ and others for types of running, etc.
Of the Manner verbs in our (C) sample (32 occurrences), the verb alha- ‘walk’ (“default general” in Matsumoto and Akita’s terminology) was the most frequent (14 of 32). 12 instances of 32 are “general nonspecific”, güi- ‘run’, üsre- ‘jump’, harai- ‘jump’, nis- ‘fly’, and höv- ‘float’, and 6 instances may be classified as Akita and Matsumoto’s “conventional specific” manner verbs, in particular, davhi- ‘gallop’, shogshi- ‘trot’, oij buu- ‘go bouncing’, and dogonc- ‘jump on one leg’.
We have examined 100 instances of motion descriptions of our (C) sample in terms of combinatorial properties. Of these, in 32 instances the ideophone modifies a manner verb (mostly with general meaning güi- ‘run’, alha- ‘walk’, üsre- ‘jump’, etc., and three instances with more specific meanings (davhi- ‘gallop’ or gete- ‘tread lightly, creep’, etc.). In 37 instances, the ideophone modifies a verb of general motion (‘go’), in 15 instances the ideophone modifies a path verb, and in 3 instances the ideophone modifies a deictic verb. In the rest (13 instances), the ideophone forms the core predicate. For specific combinations of semantic elements, see Table 1 (the core predicate element is underlined). Mongolian ideophones are therefore used either instead of, or together with manner and/or path verbs as their modifiers, see example (22).






When modifying verbs other than manner verbs, ideophones express manner. When modifying general manner verbs, they usually express subtle nuances of the motion events (see Toratani (2012: 125) for the same conclusion in Japanese). When modifying “conventional specific” manner verbs, the subtlety of nuances is even greater, and in some instances, they may only be adding expressivity to the description.
Matsumoto and Badema (2025) have convincingly shown that the Khorchin dialect of Mongolian has a high manner salience while typically using head path-coding. Among our (C) sample of 100 motion descriptions similar conclusion can be postulated. It seems that ideophones, rather than compensating for low manner salience (Slobin, 2006; Akita and Matsumoto, 2020), can be regarded as a means for the expression of narrowly specified distinctions in manner. This mainly happens in certain domains related to the traditional way of life (e.g., livestock breeding – тонтгоноод давхи- / tontgonood davhi- ‘to gallop tossing one’s head’ (of a horse), and оодгоноод давхи- / oodgonood davhi- ‘to gallop with one’s tail stretched and raised’ (of cattle).
Additionally, ideophones also take part in the semantic phenomenon of converb chaining (or stacking). Converbs in Mongolian, “nonfinite verbal form(s) that modify a following verb in the capacity of an adverbial” (Janhunen, 2012: 163), are always in overt syntactic relationship of subordination to the following verb (Janhunen, 2012:163). However, cases of “stacking” of converbs with full lexical meaning can also be interpreted as semantic coordination. These are often instances of a phenomenon described as lexical pairing (e.g., Vacek and Pürev-Očir, 1987) or binomy and apposition (Janhunen, 2012: 190–191). Examples of such “stacking” of ideophones are (20) and (21) above. The same holds also for the semantic coordination of an ideophone and a manner verb, i.e., ideophones not only modify or substitute manner verbs, but they can be combined with them to achieve new meanings.
4.4 Final Notes
As shown above, the situation of ideophones in relation to motion descriptions is rather complex in Khalkha. Rather surprisingly, given the relative richness of ideophones (ca. 700 ideophonic roots and derivational morphology, which increases the number of individual forms to thousands) as well as their readiness to appear in motion descriptions, it has turned out that roots with motional semantics are essentially lacking. On the other hand, motional semantics (in the broader sense defined by Ibarretxe-Antuñano (2019: 130–145) can be added to many ideophonic roots through verbalizing suffixes or quotative constructions.
The abovementioned (2.1) three cases of languages with rich ideophone-specific derivational morphology, in which, similarly to Mongolian, motion is coded primarily by affixes, raise a question about the comparative concept of an ideophone and the establishment of typological hierarchies (e.g., Dingemanse, 2012: 663). Since ideophones are intrinsically based on the sound-symbolic component(s), the question is whether the typological criteria delimiting ideophonic concepts in individual languages should be applied only to those non-arbitrary formal components, or (in some morphologically rich languages) to the ideophone as a whole consisting of both arbitrary and non-arbitrary morphological forms. It appears to us that the comparative concept of an ideophone should be based solely on non-arbitrary components, i.e., roots and potential expressive morphology (above all, reduplication). For the purpose of typological comparison, arbitrary derivational and inflectional forms that convey additional functions, such as affixes with motion semantics, are of secondary importance.
Any classification of the semantics of ideophones is complicated by the subtle nuances and complexity of the meanings they express. Nevertheless, we believe that by applying our set of elicitation methods, we have captured the most salient features of the semantics of the Khalkha Mongolian ideophones in motion descriptions.
5 Conclusion and Implications
Mongolian ideophonic roots primarily code the meanings of visual perceptions (shape + consistency + balance; properties of surface, properties of visibility, multiple object ordering), auditory meanings, and tactile meanings. There are no ideophonic roots whose original semantics would be exclusively motional. However, several stems gain such meanings when the roots are combined with derivative suffixes. Mongolian ideophones that appear in motion descriptions express shape in motion, modifications of manner (posture, liveliness, or lack thereof, etc.), trajectory, and speed. We have identified four groups of ideophones in terms of the degree of motional semantics’ lexicalization. The four groups appear to form a scale ranging from ideophones referring to shapes (which assume motional meanings only when followed by path verbs), to ideophones with motional ‘speed’ semantics. Between the two ends of the scale stand ideophones expressing changing posture (positioning) and ideophones describing trajectory. There are differences between the individual semantic groups of ideophones in their ability to combine with different derivational suffixes.
Regarding their syntactic properties, the data we have gathered show that ideophones can combine with all types of verbs (deictic, path, manner, etc.); they mostly occupy a peripheral position in motion constructions; however, in the form of ideophonic verbs, they can also form the core of the predicate. In such a case, they are never modified by verbs, and they are often preceded by a satellite (encoding path). Further, applying Talmy’s (2000) classification of Co-events, we have concluded that, unlike Japanese (Toratani, 2012), the Co-event of Precursion may be expressed by Khalkha ideophones.
The ability of Mongolian ideophones to appear with different types of verbs also has implications for manner salience. While Mongolian ideophones contribute to manner salience to a certain degree, the situation is different from typical verb-framed languages as described in Slobin (2006). In Mongolian, ideophones add specific features such as expressivity to motion descriptions, whether they are accompanying a manner verb or not. While ideophones per se add the highest degree of expressivity and iconic depiction, ideophonic verbs contribute these qualities to a slightly lesser degree. These results show that Mongolian is a language that can provide interesting insights into the study of the usage of ideophones in motion events (partly for its specific position with respect to Talmy’s 2000 categorization), and of ideophones in general. Our analysis of Khalkha Mongolian also generally supports all three hypotheses of Akita (2017) on encoding manner of different granularity.
Abbreviations
| abl | ablative |
| acc | accusative |
| adj | adjectivizer |
| caus | causative |
| cvb | converb |
| dat | dative-locative |
| ev | evidential |
| gen | genitive |
| ideo | ideophone |
| ins | instrumental |
| int | intensive |
| lat | lative |
| npst | non-past |
| pl | plural |
| poss | possessive |
| pres | present |
| pst | past tense |
| q | question |
| refl | reflexive |
| rpt | repetitive |
| sg | singular |
| stat | stative |
| vbz | verbalizer |
Acknowledgments
This research was supported by the Czech Science Foundation (GAČR) grant GA22-20728S, Ideophones as Bearers of Traditional Culture in Central and North-eastern Asia: Documentation and Linguistic Analysis.
References
Akita, Kimi. 2017. The typology of manner expressions: Cognitive foundations of language structure and use. In I. Ibarretxe-Antuñano (ed.), Motion and Space across Languages: Theory and Applications, 39–60. Amsterdam/Philadelphia: John Benjamins.
Akita, Kimi and Yo Matsumoto. 2020. A fine-grained analysis of manner salience: Experimental evidence from Japanese and English. In Y. Matsumoto and K. Kawachi (eds.), Broader Perspectives on Motion Event Descriptions, 143–179. Amsterdam/Philadelphia: John Benjamins.
Ameka, Felix. K. 2001. Ideophones and the nature of the adjective word class in Ewe. In F. K. E. Voeltz and C. Kilian-Hatz (eds.), Ideophones, 25–48. Amsterdam/Philadelphia: John Benjamins.
Bold, Luvsandorj et al. 2008. Mongol helnii delgerengüi tailbar tol´ (Comprehensive dictionary of Mongolian language). I-V. Ulaanbaatar: Soyombo Printing.
Brosig, Benjamin. 2014. Aspect, Evidentiality and Tense in Mongolian. From Middle Mongol to Khalkha and Khorchin. Stockholm: Stockholm University.
Byambasan, Pagma. 1970. Orchin cagiin mongol helnii üil ügiin hev, baidal. (Verbal voice and aspect in modern Mongolian). In Studia linguae et litterarum, Tomus viii, Ulaanbaatar, 201–300.
Cevel, Yadamjav. 1966. Mongol helnii tovch tailbar tol´ (The concise explanatory dictionary of Mongolian). Ulaanbaatar: Ulsyn hevleliin hereg erhleh horoo.
Cevel, Yadamjav. 2013. Mongol helnii tovch tailbar tol´. Nemj delgerüülsen hoyordugaar hevlel (The concise explanatory dictionary of Mongolian. The expanded 2nd edition). Ulaanbaatar: Mönhiin üseg.
Childs, G. Tucker. 1994. African ideophones. In L. Hinton et al. (eds.), Sound Symbolism, 178–209. Cambridge: Cambridge University Press.
Dingemanse, Mark. 2012. Advances in the cross‐linguistic study of ideophones. Language and Linguistics Compass 6(10), 654–672.
Ibarretxe-Antuñano, Iraide. 2019. Towards a semantic typological classification of motion ideophones: The motion Semantic Grid. In K. Akita and P. Pardeshi (eds.), Ideophones, Mimetics, and Expressives, 137–166. Amsterdam/Philadelphia: John Benjamins.
Jacques, Guillaume. 2013. Ideophones in Japhug Rgyalrong. Anthropological Linguistics 55(3), 256–287.
Janhunen, Juha. 2012. Mongolian. Amsterdam/Philadelphia: John Benjamins.
Kharitonov, Luka N. 1982. Grammatika sovremennogo jakutskogo literaturnogo jazyka. Tom 1: Fonetika i morfologija. (Grammar of contemporary Yakut literary language. Vol. 1: Phonetics and morphology). Moskva: Nauka.
Kullmann, Rita and Dandii-Yadam Tserenpil. 1996. Mongolian Grammar. Ulaanbaatar: Academy of Sciences, Institute of Language and Literature & Hong Kong: Jensco Ltd.
Luvsanvandan, Shadav. 1968. Orchin cagiin mongol helnii bütec. Mongol helnii üg, nöhcöl hoyor n´ (The structure of Modern Mongolian. Mongolian words and suffixes). Ulaanbaatar: ShUA-iin hevlel.
Matsumoto, Yo and Badema. 2025. Motion event descriptions in Khorchin Mongolian. In Y. Matsumoto (ed.), Case Studies of Linguistic Representations of Motion, 477–506. Berlin: De Gruyter Mouton.
Matsumoto, Yo and Kazuhiro Kawachi. 2020. Motion event descriptions in broader perspective (Introduction). In Y. Matsumoto and K. Kawachi (eds.), Broader Perspectives on Motion Event Descriptions, 1–22. Amsterdam/Philadelphia: John Benjamins.
Narantuya, Shagdarjav. 2021. Dürsleh ügiin tailbar tol´ (The explanatory dictionary of ideophones). Ulaabaatar: Udam soyol.
Nuckolls, Janis B. 2014. Ideophones’ challenges for typological linguistics: The case of Pastaza Quichua. Pragmatics and Society 5(3), 355–383.
Oberfalzerová, Alena. 2009. Onomatopoeia and iconopoeia – as an expressive means in Mongolian. Mongolo-Tibetica Pragensia 2(1), 29–60.
Oberfalzerová, Alena. 2010, The use of iconopoeic words in spoken Mongolian. Mongolo-Tibetica Pragensia ’10. Ethnolinguistics, Sociolinguistics, Religion and Culture (1), 7–34.
Oberfalzerová, Alena. 2011. The use of Onomatopoeic words in Spoken Mongolian. Mongolo-Tibetica Pragensia 4(1), 35–58.
Önörbayan, Cedev. 2000. Orchin cagiin mongol helnii üil üg büteh yos (Dagavryn argaar büteh yos) (Verb formation in modern Mongolian. The sufixal derivation). Ulaanbaatar.
Pakendorf, Brigitte. 2015. A comparison of copied morphemes in Sakha (Yakut) and Ėven. In F. Gardani, P. Arkadiev and N. Amiridze (eds.), Borrowed Morphology, 157–188. Berlin, München, Boston: De Gruyter Mouton.
Petrova, Taisija. 1948. Obraznye slova v nanaiskom jazyke (Image-Bearing Words in the Nanai Language], Izvestija Akademii Nauk sssr. Otdelenie Literatury i Jazyka [Bulletin of the Academy of Sciences of the ussr] 7(6), 527–536.
Poppe, Nicholas. 1937. Grammatika pis’menno-mongol’skogo jazyka (Грамматика письменно-монгольского языка – Grammar of Written Mongolian) Moskva-Leningrad: Nauka.
Poppe, Nicholas. 1951. Khalkha-mongolische Grammatik, mit Bibliographie, Sprachproben und Glossar. Wiesbaden, Franz Steiner Verlag GmbH.
Pürevjav, Erdene. 2010. Mongol helnii tovchilson ügiin tol´ bichig (A dictionary of Mongolian abbreviations). Ulaanbaatar: Monsudar.
Ramstedt, Gustaf I. 1908. Sravniteľnaya fonetika mongoľskago pis´mennogo jazyka i khalkhasko-urginskago govora (The Comparative Phonetics of the Written Mongolian Language and the Khalkha Dialect of Urga). S.-Peterburg.
Schaefer, Ronald P. 2001. Ideophonic Adverbs and Manner Gaps in Emai. In F. K. Voeltz, K. Erhard and Ch. Kilian-Hatz (eds.), Ideophones, 339–354. Amsterdam: John Benjamins.
Schmidt, Isaak J. 1831. Grammatik der Mongolische Sprache. St. Petersburg: Kaiserliche Akademie der Wissenschaften.
Sechenbaatar, Borjigin. 2003. The Chakhar dialect of Mongol: A morphological description. Helsinki: Suomalais-Ugrilainen Seura. Suomalais-ugrilaisen seuran toimituksia.
Slobin, Daniel I. 2006. What makes manner of motion salient. In M. Hickmann and S. Robert (eds.), Space in Languages: Linguistic Systems and Cognitive Categories, 59–81. Amsterdam: John Benjamins.
Svantesson, Jan-Olof, Anna Tsendina, Anastasia Karlsson and Vivan Franzén. 2005. The Phonology of Mongolian. Oxford: Oxford University Press.
Talmy, Leonard. 2000. Toward a Cognitive Semantics, Vol. ii: Typology and Process in Concept Structuring. Cambridge, Mass.: mit Press.
Tömörtogoo, Dom. 2018. Mongol helnii ügiin garalyn tailbar tol´ (An etymological dictionary of Mongolian language). Ulaanbaatar: Admon Print.
Toratani, Kiyoko. 2012. The role of sound-symbolic words in Motion event descriptions: The case of Japanese. Review of Cognitive Linguistics 10(1), 90–132.
Ulman, Vít. 2024. Morphosyntactic properties of ideophones in the languages of central and northeast Asia. International Journal of Eurasian Linguistics 6(1), 145–171.
Vacek, Jaroslav and B. Pürev-Očir. 1987. Lexical pairs in Mongolian (Some Theoretical and Practical Problems of Formation and Interpretation). Archív Orientální 55(4), 370–386.
Winter, Bodo, Marcus Perlman, Lynn K. Perry and Gary Lupyan. 2017. Which words are most iconic? Interaction and iconicity in the evolution of language. Interaction Studies 18(3), 443–464.
Zikmundová, Veronika. 2017. On the Mongolian Verb of Motion yav- ‘to Go, to Travel, to Leave.’ In B. Khabtagaeva and S. Szeverényi (eds.), Uralic and Siberian Lexicology and Lexicography, 51–74. Szeged.
Mongolian has several major dialects, sometimes classified as separate languages, in addition several Mongolian literary languages are used (in particular Written (Literary) Mongol, and Cyrillic Khalkha).
We list Talmy’s (2000) structural concepts according to accepted convention with an initial capital letter.
In this paper we use both Mongolian Cyrillic and its common transcription in Latin adjusted for use in scientific papers (by distinguishing ү > ü and ө > ö and use of ц > c instead of ts) following Brosig (2014: 173).
Capital A in suffixes indicates that the suffix has four vocalic forms distributed according to the rules of vowel harmony following the vowels of the stem. The four vocalic variants are a, e, o and ö.
When the ‘do’ or ‘say’ verb is followed by another verb and therefore takes the form of a converb, the simultaneous and consequent converb are used interchangeably.
The questionnaires and the motion description sample are available at <https://doi.org/10.5281/zenodo.15526830>.
The Khalkha Mongolian verb yav- ‘to go’ has two basic meanings – that of general motion (‘go’, ‘travel’, ‘drive’, etc.) and that of deictic verb (‘leave’), cf. Matsumoto and Badema (2025); Zikmundová (2017).
