Part of speech tagging – Corpus of Historical Low German

Since the Middle Low German component of the CHLG is a collaboration with the Referenzkorpus Mittelniederdeutsch / Niederrheinisch (1200-1650) (ReN), the part of speech tagging uses the HiNTs tagset (Barteld et al. 2018).

Compared to the standard and rather broad tagset intended in the original Penn annotation scheme, the HiNTs tagset involves a much more fine-grained set of distinctions. The full documentation of how the HiNTS tagset was employed for the ReN corpus can be found here (Annotationshandbuch Part 1, in German).

The HINTS tags employs recurring labels which encode certain values for the features of lexical identity, articleness and position/headedness, as follows:

Feature	Value	Label within the tag
lexical identity	definite	`D`
	indefinite	`I`
	negative	`NEG`
	possessive	`POS`
	relative	`REL`
	wh	`W`
articleness	article-like	`ART`
articleness	not article-like	–
position/headedness	pre-head	`A`
	post-head	`N`
	head	`S`
	head of nominal predicate	`D`

These labels are combined to produce individual POS-tags, for example:

DDARTA: a determiner which is definite, article-like and precedes its head noun, e.g. dat land ‘that land’
DIARTA: a determiner which is indefinite, article-like and precedes its head noun, e.g. eyn kind ‘a child’
ADJA: an adjective which precedes its head noun, e.g. grot wunder ‘great wonder’
ADJN: an adjective which follows its head noun, e.g. juncvrowe fin ‘maiden fine‘
ADJS: an adjective which is a head, e.g. Se was nicht de beste ‘She was not the best‘
ADJD: an adjective which is the head of a nominal predicate, e.g. De neppe weren grot unde runt ‘the bowls were big and round‘

Here, we provide an overview of how some of the key HiNTs tags correspond to the more standard Penn tagset:

Penn tag	HiNTs tag	Comment
`ADJ`	`ADJA`	attributive adjective pre-head
	`ADJN`	attributive adjective post-head
	`ADJS`	substantival adjective as head
	`ADJD`	adjective as head of nominal predicate
`ADV`	`AVD`	adverb
	`ADJV`	adverbial adjective
	`AVKO`	conjunctional adjective
`CONJ`	`KON`	coordinating conjunction
`D`	`DDARTA`	definite article pre-head
	`DDARTN`	definite article post-head
	`DIARTA`	indefinite article pre-head
	`DIARTN`	indefinite article post-head
	`DDA`	demonstrative article pre-head
	`DDN`	demonstrative article post-head
	`DPDS`	demonstrative pronoun
`FW`	`FM`	foreign material
`INTJ`	`ITJ`	interjection
`N`	`NA`	common noun
`NPR`	`NE`	proper noun
`NEG`	`PTKNEG`	negation particle
`NUM`	`CARDA`	attributive numeral pre-head
	`CARDN`	attributive numeral post-head
	`CARDS`	substantival numeral as head
	`CARDD`	numeral as head of nominal predicate
`P`	`APPR`	preposition
`P`	`APPO`	postposition
`PRO`	`PPER`	personal pronoun
`PRO-RFL`	`PRF`	reflexive personal pronoun
`PRO$`	`DPOSA`	possessive adjective pre-head
	`DPOSN`	possessive adjective post-head
	`DPPOSS`	substantival possessive as head
	`DPOSD`	possessive as head of nominal predicate
`PTCL`	`PTKA`	particle with adjective or adverb
`PTCL`	`PTKN`	particle with nominals
`RP`	`PTKVZ`	verbal particle
`TO`	`PTKZU`	infinitival particle

Verbs are treated along two different dimensions in the HiNTs tagset:

There is a three-way distinction between lexical verbs (VV), modal verbs (VM) and HAVE, BE, BECOME and DO as auxiliaries (VA)
There is a four-way distinction finite verbs (FIN), infinites (INF), imperatives (IMP), present particles (PS) and past/passive participles (PP)

These are combined as follows:

Broad category	Tag	Comment
lexical verbs	`VVFIN`	finite
	`VVINF`	infinitive
	`VVIMP`	imperative
	`VVPS`	present participle
	`VVPP`	past/passive participle
modal verbs	`VMFIN`	finite
	`VMINF`	infinitive
	`VMIMP`	imperative
	`VMPS`	present participle
	`VMPS`	past/passive participle
auxiliaries (HAVE/BE/BECOME/DO)	`VAFIN`	finite
	`VAINF`	infinitive
	`VAIMP`	imperative
	`VAPS`	present participle
	`VAPP`	past/passive participle