Udmurt spatial cases dataset

This is the complete dataset used in my thesis he semantics of Udmurt spatial cases. The data is used to analyze the distinct senses (independent meanings) that the semantic structure of the cases consists of. It is V3.0 of Udmurt spatial cases dataset, and it is greatly extended compared to V2.0. The dataset contains an annotated set of examples on all spatial cases of Udmurt, that is inessive, elative, egressive, illative, terminative, prolative, and approximative. There are 500 examples of each case except for approximative, so the size of the complete dataset is 3100 examples. There is one file named [casename]_eng.csv.

For the V3.0 release the data analysis of the data has been checked and corrected where necessary, so the analyses do not always correspond to those in V2.0. The dataset is manually annotated for the following variables which are hypothesized to affect the senses of the cases:

Sense
The sense and contextual variant (in brackets) of the spatial case inflected word; categories as defined in Erkkilä (2026)

Conceptual domain
The conceptual domain to which the even is situated; categories as defined in Erkkilä (2026)

Whole LM phrase & translation
The phrase coding the Landmark (LM) and its literal translation

LM
Translation of the word expressing LM, except for relational nouns with a modifier, for which the modifier is translated.

LM word class
The word class of the word coding the LM
ADJ = adjective
ADV = adverb
DMST = demonstrative
DVB = deverbal noun
INDF = indefinitive
INTG = interrogative
NUM = numeral
REL = relativizer
RN-ADV = relational noun without a modifier
RN-PP = relational noun with a modifier
SUBS = noun

LM type
The type of the thing functioning as the LM
abstract = an abstract thing
amount = an amount
concrete = a concrete (physical) thing
institution = an institution
temporal = a temporal thing

Functional dimensions
How many functional dimensions does the LM have; 0 = a thing witohut dimensions (e.g., an abstract o temporal thing), 1–3 = the number of the funtional dimensions (this parameter should be used with caution, as the values have not been verified by native speakers)

Mass/multiplex entity, Plural & Possessive marking
Is the LM a mass or multiplex entity, and does the word coding the LM have plural or possessive marking (the use of these parameters should be avoided, as the coding has not been consistent)

Whole verb phrase & translation
The verb phrase that the associated verb form is part of; in translation column EXT = existential predication (no verb), ATTR = deverbal noun (no verb)

Temporal reference
The tense, modal, or non-finite marking of the verb phrase
FUT = future tense
IRR (CND) = conditional mood
IRR (IMP) = imperative mood
IRR (NEC) = non-finite necessitative form
IRR (NEG) = non-finite negative form
IRR (OPT) = optative mood
PL.PRF = pluperfect tense
PRF = perfect tense
PST = predicative past participle
PST1 = 1st past tense
PST2 = 2nd past tense
PRS = present tense

Negation
Is the verb phrase negated or not

Associated verb form
The translation of the associated verb form

Finiteness
Is the associated verb form finite or not
ATTR.PST.PTCP = attributive past participle
FIN = finite
INF = infinitive
LIM.CVB = converb of limit
MOD.CVB = modal converb
MOD.PTCP = modal participle
NEG.PTCP = negative participle
NMLZ = deverbal noun
PCL = optative particle
PRD.PST.PTCP = preddicative past participle
PRS.PTCP = present participle
SIM.CVB = converb of simultaneity

Aspect type
Aspect type of the associateed verb form; categories as defined in Croft (2001)
acquired state
directed activity
incremental accomplishment
inherent state
irreversible achievement
nonincremental accomplishment
reversible achievement
semelfactive
transitory state
undirected activity

Action type
Type of the process the associated verb form expresses
directed non-translocational = directed action wihtout translocation
directed scalar = directed action along a scale
directed translocational motion = directed motion with a change of place
directed translocational non-motion = directed action (not motion) with a change of place
non-directional state = state
undirected non-translocational = action without a direction or motion
undirected translocational motion = motion with change of place but without a clear direction
undirected translocational non-motion = action (not motion) with change of place but without a clear direction
unspecified non-translocational = motion without a change of place which can be understood contextually as directed or undirected
unspecified translocational motion = motion with a change of place which can be understood contextually as directed or undirected

Causation & Iterativity/continuativity
Does the associated verb form have a derivational suffix expressing the notion

Boundedness
Is the event bounded; categories as defined in Depraeter (1995)
beginning
bounded
end
unbounded

In addition the following data can be found in the tables:
- the corpus sentence in Udmurt in its original, but sometimes abbreviated form
- a translation in English (the translations are not always idiomatic, and thus they shoukd not be used as anything else than to get a rough idea of what the Udmurt senstences mean)
- the original source of the sentence as given in the corpus with the following abbreviations (the numbers give day, month, and/or year of the publication of the original source as given in the corpus)
A = Aźlane (Häzmät)
anekdotjos = Анекдотъёс (serem.ru)
annushka = annushkazagumennova.blogspot.ru Анна Загумённова
Apostoljos = Апостолъёслэн ужъёссы М. Г. Атаманов (берыктон)
Arxipov = Кинъёс соос? (Трофим Архипов)
Avdij = Авдий пророклэн книгаеза М. Г. Атаманов (берыктон)
Avvakum = Аввакум пророклэн книгаез М. Г. Атаманов (берыктон)
B_Wiki = Бесерманъёс Wikipedia
BN = Байгурезь (Новый путь)
burdjos = burdjos.ru Мария Золотарёва
chimali = chimali.ru (Ксения Ворончихина)
chinkudi = chinkudi.blogspot.ru Ксения Евсеева
chudja = chudja-zheni.ru (Евгений Бикузин)
DI = Дом дружбы народов - Иворъёс
DZ = Ӟечбур!
Eksejlyk = Эксэйлыкъёслэн куинетӥ книгазы М. Г. Атаманов (берыктон)
Ermolaev = Алексей Ермолаев. Известный и незнаемый А. А. Ермолаев, Н. Ермолаева (составитель) 1959–2009
I = Idnakar
I_Wiki = Иванова, Маргарита Григорьевна Wikipedia
Ieremija = Иеремиялэн бӧрдонэз М. Г. Атаманов (берыктон)
igonina = igoninamaria.ru Мария Игонина
inzhelka = inzhelka.tumblr.com Елизавета Шихова
Ioann = Иоаннлэсь ӟеч ивор М. Г. Атаманов (берыктон)
Ioann_usjtisjkon = Иоанн апостоллы усьтӥськон М. Г. Атаманов (берыктон)
IU = IA "Udmurtija"
Ja_Wiki = Яр ёрос Wikipedia
Jashin = Гондыр но лудкеч (Даниил Яшин, Роза Яшина)
Jozhev = Чылкыт сюлэмын Аттӥла Йожеф
Ju_Wiki = Юкаменск ёрос Wikipedia
K_Wiki = Киров карлэн историез Wikipedia
kepics = kepics.tumblr.com Ольга Игнатьева
KM_Wiki = Кедра Митрей Wikipedia
kuala = kualaluly.ru Елена Ильина
kyshman = kyshman.blogspot.ru Ксения Кадрова
kyshno = kyshnomurt.ru
ladamiri = ladamiri-lari.livejournal.com Лариса Дмитриева
Luka = Лукалэсь ӟеч ивор М. Г. Атаманов (берыктон)
marjamoll = marjamoll.blogspot.ru
Mark = Марклэсь ӟеч ивор М. Г. Атаманов (берыктон)
Matfej = Матфейлэсь ӟеч ивор М. Г. Атаманов (берыктон)
MU = Mynam Udmurtije
muketulon = muketulon.tumblr.com Алина Крестьянинова
N_Wiki = Насибуллин, Риф Шакрисламович Wikipedia
nylkyshno = nylkyshnokenesh.blogspot.ru Удмурт нылкышно кенеш
O = Ošmes
OE = Ошмес (Эстониысь удмуртъёслэн газетсы)
OR = Ошмес (Рассвет)
Pavel_Efes = Павел апостоллэн Ефес карын улӥсьёслы ыстэм гожтэтэз М. Г. Атаманов (берыктон)
Pavel_gozhtet = Павел апостоллэн Тимофейлы ыстэм кыкетӥ гожтэтэз М. Г. Атаманов (берыктон)
Pavel_Korinf = Павел апостоллэн Коринф карын улӥсьёслы ыстэм кыкетӥ гожтэтэз М. Г. Атаманов (берыктон)
Pavel_Rim = Павел апостоллэн Рим карын улӥсьёслы ыстэм гожтэтэз М. Г. Атаманов (берыктон)
Pojarkova = Киям крезе, сюлмам - кырзанэ Т. А. Пояркова, Т. П. Четкарёва
pushdun = pushdun.blogspot.ru (Миклай Михайлов)
Raud = Сипсик Эно Рауд
Sadikov = Р. Р. Садиков: Кисса гурт: вашкала но туала (гурт калыклэн историез но этнографиез)
sanko = sankobatyr.blogspot.ru
Sh_Wiki = Шор ёросысь удмуртъёслэн дӥськутсы Wikipedia
shushynati = shushynati.ru Наталья Михайлова
SP = Светлый путь
SP_Wiki = Санкт-Петербург Wikipedia
Tarasova = Пейзажной лирикаез азинтон удысын Гердлэн тыршемез (Ежегодник финно-угорских исследований, 2014, специальный выпуск) А. Н. Тарасова
tubat = tubat.blogspot.ru (Мария Турова)
U_Wiki = Удмуртия Wikipedia
UD = Udmurt dunne
udmurto = udmurto4ka.blogspot.ru
UK = Удмурт элькунлэн Кун кенешез
V_Wiki = Васильев, Флор Иванович Wikipedia
vamentul = vamentul.blogspot.ru (Наталья Образцова)
Vanjushev = Герд улэ асьме пӧлын (Ежегодник финно-угорских исследований, 2014, специальный выпуск) В. М. Ванюшев
VD = Vańmon dyr
vorshud = Шкляев Александр vorshud.unatlib.ru
vuflower = vuflower.blogspot.ru Лилия Кондратьева
Z_Wiki = Завьял ёрос Wikipedia
zangarik = zangarik.ru (Ирина Зангарик)
Zh_Wiki = Жириновский, Владимир Вольфович Wikipedia

A hyphen (-) in a cell means that the parameter is not applicable to the example in question.

The V3.0 release also includes the V1.0 and V2.0 releases. The V1.0 and V2.0 contain a dataset (the files wiht names starting Goal or Source) and accompanying R code (the files with names starting with R) for studying the variation between the two source cases (elative and egressive) and goal cases (illative and terminative). The data folder include four files including data for analysing the effect of properties of the Landmark and the associated verb form on the choice between the source and goal cases. The script folder contains four R markdown files, which contain the code to analyse the four data files.

The V2.0 dataset is annotated for various variables that are hypotethised to affect the choice between the pairs of cases. Not all the variables are used in the analysis presented. The variables area subset of the variables of V3.0, and they are coded differently. The correspondences in coding between V3.0 and V2.0 are as follows:

Temporal reference
FUT = FUT
INF, IRR (CND), IRR (IMP), IRR (NEC), IRR (NEG), IRR (OPT) = IRR
PL.PRF, PRF, PST, PST1, PST2 = PST
PRS = PRS

Aspect type
acquired state = sta_acq
directed activity = act_dir
incremental accomplishment = acc_inc
inherent state = sta_inh
irreversible achievement = ach_nrev
nonincremental accomplishment = acc_ninc
reversible achievement = ach_rev
semelfactive = ach_sem
transitory state = sta_tra
undirected activity = act_ndir

Action type
directed non-translocational = dir_ntra
directed scalar = dir_sca
directed translocational motion = dir_tra_mo
directed translocational non-motion = dir_tra_nmo
non-directional state = ndir_sta
undirected non-translocational = ndir_ntra
undirected translocational motion = ndir_tra_mo
undirected translocational non-motion = ndir_tra_nmo
unspecified non-translocational = nsp_ntra
unspecified translocational motion = nsp_tra_mo

Boundedness
beginning, unbounded = nend (with goal cases)
bounded, end = end (with goal cases)
beginning, bounded = start (with source cases)
end, unbounded = nstart (with source cases)

In V2.0 (compared to V1.0) the R code is heavily changed. Only the code in files R_goal3 and R_source3 should be used in analysis. The data files are the same in both versions, but in V2.0 they have been checked and some minor mistakes have been fixed. Therefore, only the data files in the new version should be used if replicationg the analysis represented by the code in files R_goal3 and R_source3.

The author of this dataset is Riku Erkkilä and it is published under CC-BY-NC-ND licence. If used in a publication, please refer to this publication as well as mention the original source:

Arkhangelskiy, Timofey 2018: Udmurt corpus. http://udmurt.web-corpora.net/index.html.

The R packages used in the analysis are:
Hothorn, Torsten; Hornik, Kurt; Strobl, Carolin & Zeileis, Achim 2023: Party 1.3-13. https://CRAN.R-project.org/package=party.
Wickham, Hadley 2023: Tidyverse 2.0.0. https://CRAN.R-project.org/package=tidyverse.

The V2.0 dataset is used in the following publications:
Erkkilä, Riku 2024: The interplay of conceptualization and case marking in the directional cases of Udmurt. Yearbook of the German cognitive linguistics association 12. https://doi.org/10.1515/gcla-2024-0006.

The V3.0 dataset is used in the following publications:
Erkkilä, Riku 2026: The semantics of the Udmurt spatial cases. Helsinki: University of Helsinki.