The tables on this page list the tags used by the IPCHG and their meanings. Our tags are based on the Penn family of parsed historical corpora and are most similar to those in HeliPaD. The tables below list the PPCHE and HeliPaD equivalents of our tags, when these differ from ours. (A blank cell indicates that the corpora use the same tags; n/a means that the corpus does not have a tag for this category. In addition, there are a small number of tags used by other corpora that we do not use: ALSO, DO, ELSE, FP, GE, HAN, OTHER, SUCH, etc.)

Our source corpora are part-of-speech tagged using the Stuttgart-Tübinger Tagset (STTS) or the related Historisches Tagset (HiTs), which differ substantially from the Penn-type tagsets. An important difference is that STTS/HiTs tags include some syntactic information, which are not reflected on the heads in our system but are taken into consideration when we determine structure. For example, adjectives are ADJA in STTS when attributive preceding the N, ADJN when attributive following the N, ADJS when substantivized, and ADJD when predicative; in our corpus, these are all ADJ and the differences are captured by the appearance of the adjective in an NP with a noun, without a noun, in a predicative ADJP, etc, e.g. As another example, we tag the past participle of a lexical verb as VBN whether it is used verbally, adjectivally, or substantivized; the difference between these uses can be seen in its syntactic position (constituent of the clause vs. ADJP vs. NP).

Inflected parts of speech are extended with morphological tags.

Morphological extensions

As in HeliPaD, we follow a maximalist strategy, tagging all relevant heads for case, number, and person, whether or not inflection is overt. The purpose is not to create a fully morphologically tagged corpus, but rather to allow researchers to search for syntax/morphology interactions that may be of interest. We tag based on form (when clear) and/or semantics/context (relying on e.g. dictionaries like Lexer); in truly ambiguous instances we select SG>PL, N>A>D>G, and indicative>subjunctive.

Nouns and their modifiers (including participles) are labelled for case and number with the extensions ^N^SG, ^A^SG, ^D^SG, ^G^PL, etc. (In HeliPaD, pronouns are additionally marked for person, which we have not (yet) implemented.) The only nominal heads that are not case/number tagged are predicative adjectives (and by extension, participles) in ADJP-PRD.

The case/number of possessive pronouns reflects the case/number of the noun that it modifies, e.g. unserem is tagged as PRO$^D^SG (despite the fact that unser refers to a plural possessor).

Finite verbs, in addition to the tense/mood tagging included in the main part of speech tag, are labelled for person and number, e.g. ^1^SG, ^3^PL, etc.