about data annotation queries sources citation

Pontic Greek

is a dialect of Greek, originally spoken in various regions of northeastern Anatolia. It is often regarded as a conservative variety of Greek, preserving many features of Middle Greek. From the 18th to 20th century, speakers of Pontic Greek migrated from Anatolia to the Caucasus in several waves.

Contributors

Svetlana Berikashvili

recordings, transcriptions, translations, glossing, analyses

Ilia State University

Stefanie Böhm

instructions, revisions

Bielefeld University

Evgenia Kotanidis

recordings, transcriptions, translations

Ivane Javakishvili Tbilisi State University

Johanna Lorenz

instructions, revisions

Bielefeld University

Stavros Skopeteas

conception, supervision, (few) recordings

University of Göttingen

Language

Genetic affiliation: Indo-European > Greco-Phrygian > Greek > East Greek > Koineic Greek > Pontic-Cappadocian Greek > Pontic (Glottolog 4.0)
Place: Georgia
Language Code: pnt (ISO 639-3)
Population: below 500
Documented variety: Pontic Greek of Georgia, as spoken by Pontic Greek people in Georgia, and from speakers migrated from there to Greece.
Endangerment: threatened (Endangered Languages Project)

Glottolog 4.8 edited by Hammarström, Harald & Forkel, Robert & Haspelmath, Martin & Bank, Sebastian, licensed under a Creative Commons Attribution 4.0 International License.

data

Time and Place

Most recordings were made between 2014 and 2016, with a small number dating back to 2005.

Data were collected in Georgia (Tsikhisjvari, Manglisi, Santa, Shua Kharaba, Tetritskaro, Tbilisi) and in Greece (Athens and Thessaloniki, from Pontic speakers who had emigrated from Georgia). The map shows the recording locations in Georgia as well as the places of birth of the speakers in our sample.

Speakers

subcollection TXT, women = 13, men = 11, Birthyear range = 1925-1994, average = 1966,5
subcollection VA1, women = 4, Birthyear range = 1933-1981, average = 1955,5
subcollection VA2, women = 1, men = 1, Birthyear range = 1930-1940, average = 1935

Instructions

The speakers were instructed in Pontic Greek by a native instructor. The instructions (TXT, VA1 subcollections) are listed below in English translation.

Abbreviation	Text	Instruction
	General instruction	Please answer the following questions spontaneously. Just speak normally, as if you are speaking to a friend. It does not matter if you are not sure about details, just give a natural answer.
AN	Ancestor Story	How did your ancestors come to Georgia?
FM	Family	Tell us the history of your family (for speakers who do not live in the original settlements: how did your family came from the villages to Tbilisi and from Tbilisi to further destinations)?
VL	Village	Please describe the village where your family comes from.
CD	Comparative Description	Please tell us how your people are different from the other people in the village/city (Russian, Greek)?
CL	Culture	Please tell me a fairy tail or a poem in your native language. (If you do not know any fairy tail/poem, please tell me what you find most important in the culture of your people).
MR	Marriage	Please tell us how your people celebrate an engagement/marriage and what is the difference to the way other people in this village/city feel celebrate a marriage.
FE	Feast	Tell us a difference between the way you celebrate a particular feast in your group and the groups of the other people of your environment? (Christmas, Easter, Panajia).
LG	Language	Please tell me how you perceive the major differences between your language and Russian.

Recordings

Recordings took generally place in quite places (not laboratory conditions).
Recordings were made with an Olympus LS-11, in-built microphones, at a sampling frequency of 44.1 kHz, audio files were saved in wav format.
subcollection TXT, soundfiles = 338; duration = 22030 sec (6.1 h); words = 35843
subcollection VA1, soundfiles = 30; duration = 2278 sec (0.6 h); words = 3830
subcollection VA2, soundfiles = 5; duration = 2029 sec (0.56 h); words = 3400

File names

Labelling scheme: language-collection-instruction-00000-speaker

Example: PNT-TXT-AN-00000-A01
PNT- language code: PNT=“Pontic Greek" (ISO-639-3)
TXT- subcollection code: TXT=“Texts" (other options: VA1, VA2)
AN- instruction "Ancestors" (other options: see instructions above)
00000- does not play a role in this subcollection
A01 Speaker identifier, sample A: original settlements, B: internal migration, C: emmigration to Greece

Example: PNT-TXT-CL-2-000-A04
PNT- language code: PNT=“Pontic Greek" (ISO-639-3)
TXT- subcollection code: TXT=“Texts" (other options: VA1, VA2)
CL- instruction "Culture" (other options: see instructions above)
2- second text of the same speaker with the same instruction
000- does not play a role in this subcollection
A04 Speaker identifier, sample A: original settlements, B: internal migration, C: emmigration to Greece

Annotations

Layers

The annotation layers of the corpus files (.eaf) contain the following layers:

Tier name	Content
ref	identifier of the sentence: contains the name of the text and the number of the sentence in the last two digits; it provides the reference to cite examples and to store references in searching results
tx	PARENT=ref corrected transcription on word level: makes it possible to search for single words containing particular glosses, that would be not possible if only the tx-a tier would exist; besides, the word boundaries can be used later to align word boundaries and sound
mb	PARENT=tx corrected transcription on morpheme level
ge	PARENT=mb morpheme-aligned glosses in English
ps	PARENT=mb parts of speech
ft	PARENT=ref free translation
nt	PARENT=ref comments / notes about the sentence
tx_a	PARENT=ref transcription on sentence level: this tier makes it possible to search for discontinuous strings of words hosted by a single ref-entry and with an undefined sequence of words in between; it also facilitates reading the whole sentence in the object language
ge_a	PARENT=ref (a = associated) sentence-aligned glosses in English: this tier makes it possible to search for discontinuous strings hosted by a single ref-entry and with an undefined sequence of words in between at the functional level (not at the level of forms, unlike tx_a)
id	no content; needed for back-conversion to Toolbox

In files with more than one speakers, the speaker label is merged to the content label (this applies to subcollections VA1 and VA2):

...@speaker1
...@speaker2
...@speaker3

Transcriptions

class	orthography	IPA
vowels	a	a
	ä	æ
	e	ɛ
	i	i
	o	ɔ
	u	u
plosives	p	p
	t	t
	k	k/c
	b	b
	d	d
	g	g/ɟ
fricatives	f	f
	θ	θ
	s	s
	sh	ʃ
	x	χ
	v	v
	ð	ð
	z	z
	zh	ʒ
	j	j
	ɣ	ɣ
affricates	ts	ʧ
	ch	ʧ
	dz	dz
	dzh	ʤ
nasals	m	m
	n	n/ŋ
liquids	r	ɾ
	l	l/ł

Notes:

[ŋg] is transcribed as consonant cluster, e.g., angelos 'angel'
[pʃ] is transcribed as consonant cluster, e.g., pshi 'soul'
We transcribe lemmata – not phonetic instantiations: Some variation in the consonant realization may depend on the influence of Georgian; a case at issue is the glottalization of plosives: k’ – k, p’ – p, t’ – t (following the Georgian national system of Romanization). The glottal versions are more frequent in case of k’. /paliá/ is sometimes realized with a plosive [p], sometimes with a glottalized [p’]; /etílien/ is sometimes realized with a plosive [t], sometimes with a glottalized [t’]; /k/ is sometimes realized with a plosive [k], sometimes with a glottalized [k’], e.g., in /kéxume/
Allophonic variation:/k/ and /g/ are realized have palatal allophones [c] and [ɟ] following front vowels (/i/, /e/, /ä/, and velar allophones elsewhere.
Variation between ia, ja and æ: synizesis does not always applies in final diphthongs -ia e.g. karðía ‘heart’, in such cases transcription is given with ia; semivowels, e.g. maimúnja ‘monkeys’, turkjá ‘Turks’, kanónja ‘canons’ then it is transcribed with -ja. The spelling -ja is used for lemmata such as jatí ‘why’, jajá ‘grandmother’ with another value; the -ä, which is close to the spelling, is used to represent IPA value [æ]. It is mainly the result of contraction of and , when they are unstressed or when the stress falls on the second vowel, e.g. kartófä ‘potatoes’, ðävéno ‘pass’, tämizljávo ‘clean’ etc. (For the detailed discussion and examples see Tombaidis 1988: 33-34, Revithiadou & Spyropoulos 2009: 41-43; 2012: 30-31)

Category	Abbreviation	Meaning
Gloss	0	epenthesis
Gloss	1	first person
Gloss	2	second person
Gloss	3	third person
Gloss	ABIL	ability
Gloss	ABL	ablative
Gloss	ACC	accusative
Gloss	ADJR	adjectivalizer
Gloss	AOR	aorist
Gloss	COND	conditional
Gloss	COND.COP	conditional copula
Gloss	CONV	converb
Gloss	DAT	dative, genitive
Gloss	EPST.COP	epistemic copula
Gloss	EV.PST	evidential past
Gloss	FUT	future
Gloss	GEN	genitive
Gloss	GER	gerund
Gloss	INF	infinitive
Gloss	INSTR	instrumental
Gloss	IPFV	imperfective
Gloss	NEG	negative
Gloss	NEG.COP	negative copula
Gloss	NEG.EXIST	negative existential
Gloss	NR	nominalizer
Gloss	OPT	optative
Gloss	PASS	passive
Gloss	PL	plural
Gloss	POSS	possessive
Gloss	POT	potential
Gloss	PROC	procedural
Gloss	PST	past
Gloss	PTCP	participle
Gloss	SG	singular
Gloss	VOC	vocative
Part of speech	N	noun
Part of speech	V	verb
Part of speech	A	adjective
Part of speech	Adv	adverb
Part of speech	P	adposition
Part of speech	Q	quantifier
Part of speech	AQ	ordinal nominal
Part of speech	C	conjunction
Part of speech	PN	pronoun
Part of speech	PRT	particle
Part of speech	X	unclear
Miscellaneous	xxx	unidentified words (mb layer)
Miscellaneous	xxx	unknown meaning (ge layer)
Miscellaneous	HESIT	hesitation
Miscellaneous	((coughs))	coughing
Miscellaneous	((laughs))	laughing
Miscellaneous	((smiles))	smiling

queries

The Pontic Greek corpus is online available in ANNIS that allows for visualizations and queries in multimodal annotations. It comes with a powerful query language (AQL=ANNIS query language) that allows to retrieve complex data patterns in multilayered annotations (Krause, Thomas & Zeldes, Amir 2016: ANNIS3: A new architecture for generic corpus query and visualization. in: Digital Scholarship in the Humanities 2016 (31). http://dsh.oxfordjournals.org/content/31/1/118).

You can access the corpus in the SPW installation at: https://spw.uni-goettingen.de/annis/.

Before the query, you need to select a corpus: e.g., "PNT-TXT-0.1-mp3". You write your query in the query window:

Plain-text queries in AQL

In your queries, you need to specify the annotation layer and define the expression that you are looking for. Notice that the query tool only retrieves sentences that equal the queried expression (and not sentences that contain the queried expression).

Query	Explanation
tok	all tokens of the corpus (not very useful, but illustrative)
mb="ke"	all tokens in the morpheme layer (mb: words with morphemic boundaries) that contain exactly the form ke 'and'.
mb="aðaká"	all tokens in the morpheme layer (mb) that contain exactly the form "aðaká".
ge="be:PST:3.PL"	all tokens in the gloss layer (ge) that contain exactly the string "be:PST:3.PL".
ge="LOC"	all tokens in the gloss layer (ge) that contain exactly the string "LOC" (= locative).
ps="V"	all tokens in the POS layer (ps) that contain exactly "V" (=verb)
mb="s" _=_ ge="LOC"	all tokens that contain exactly "s" in the morpheme layer and exactly "LOC" in the gloss layer - in the same slot (_=_).

Regular expressions in AQL

Regular expressions are included in slashes. You find some illustrative examples below. More details about the regular expressions in AQL are found here.

Query	Regular expression	Explanation
mb=/eksér[oi]/	"[...]" contains alternative characters	all tokens in the morpheme layer (mb) that contain "ekséro" or "ekséri".
mb=/(pos\|pu)/	"(...\|...)" contains alternative strings	all tokens in the morpheme layer (mb) that contain "pos" or "pu".
mb=/ekséro?/	"?" stands for "the last character is optional"	all tokens in the morpheme layer (mb) that contain the string "ekséro" or "eksér".
mb=/a+/	"+" stands for "at least one occurrence"	all tokens in the morpheme layer (mb) that contain the at least one occurrence of the character "a", which includes "a" and "aa".
mb=/a*/	"*" stands for "zero or more occurrences"	all tokens in the morpheme layer (mb) that contain "a", "aa", "aaa", "aaaa", etc.
mb=/ti./	"." stands for "whatever character"	all tokens in the morpheme layer (mb) that contain the string "ti" and a character (.), e.g., "tin", "tis", etc.
ge=/.*3.PL/	".*" stands for "zero or more occurrences of whatever character"	all tokens in the gloss layer (ge) that contain "...3.PL"
ge=/.PFV./	".*" stands for "zero or more occurrences of whatever character"	all tokens in the gloss layer (ge) that contain "...PFV..."

More about AQL: AQL documentation site.

publications using this corpus

Berikashvili, Svetlana. 2022. Contact-Induced Change in the Domain of Grammatical Gender in Pontic Greek spoken in Georgia. Languages 7(2): 79. https://doi.org/ 10.3390/languages7020079.

Berikashvili, Svetlana. 2019. Verb Adaptation in Pontic Greek spoken in Georgia. In Tzitzilis, Ch. & G. Papanastassiou (eds). Language Contact in the Balkans and Asia Minor, Series: Greek Language: Synchrony and Diachrony 2. Thessaloniki: Institute of Modern Greek Studies (M. Triandaphyllidis Foundation), 262-279.

Berikashvili, Svetlana. 2018. Several Features of Aorist and Verbal System in Pontic Greek spoken in Georgia. Arxeion Pontou (Pontic Archive), Vol. 58. Athens: Committee for Pontic Studies, 195-229.

Berikashvili, Svetlana. 2017. Morphological Aspects of Pontic Greek spoken in Georgia. Series: Languages of the World 54, Munich: LINCOM 2017, 168pp.

Berikashvili, Svetlana. 2016. Morphological Integration of Russian and Turkish Nouns in Pontic Greek. Language Typology and Universals, 69.2, 255–276, DOI: 10.1515/stuf-2016-0012.

Berikashvili, Svetlana & Lobzhanidze, Irina. 2017. Number in Pontic Greek spoken in Georgia. In M. Chondrogianni, S. Courtenage, G. Horrocks, A. Arvaniti, I. Tsimpli (Eds.), Proceedings of the 13th International Conference on Greek Linguistics. London: University of Westminster, 51-61.

citations

TXT subcollection

Kotanidi, Evgenia (collection/transcription/translation), Svetlana Berikashvili (revision/glossing), Stefanie Böhm, Johanna Lorenz (supervision), Stavros Skopeteas (design/supervision). 2019. Pontic data collection. The Language Archive, Corpus resource; persistent identifier: https://hdl.handle.net/1839/00-0000-0000-0021-4DA4-3.

VA1 subcollection

Berikashvili, Svetlana. 2019. Pontic data collection 2. The Language Archive, Corpus resource; persistent identifier: https://hdl.handle.net/1839/00-0000-0000-0021-4DA4-3.

VA2 subcollection

Berikashvili, Svetlana (transcription) and Stavros Skopeteas (recordings). 2019. Pontic data collection 3. The Language Archive, Corpus resource; persistent identifier: https://hdl.handle.net/1839/00-0000-0000-0021-4DA4-3.