Documentation Version 7.0
You can find the more detailed annotation guidelines here.
Corpus pipeline
- 8 additional texts were added:
ArtzneyBuchleinDerKreutter-Abrotanum_1532_TallatYou can find a complete list of all documents of this version in the annotation guidelines.
ArtzneyBuchleinDerKreutter-Centaurea_1532_Tallat
ContrafaytKreuterbuch-CCLVII-CCLXX_1532_Brunfels
ContrafaytKreuterbuch-CCLXXI-CCXC_1532_Brunfels
GartDerGesundheit-c15-23_1487_Cuba
GartDerGesundheit-c23-33_1487_Cuba
NewKreuetterbuch-CXXXIIIJ-CXXXVIIJ_1539_Bock
NewKreuterbuch-c19-25_1543_Fuchs
- Automatic preparation of an OCR for all new documents of version 7. The documents of version 1-6 were transcibed manually.
- Tokenization of the OCR/transcription with TreeTagger.
- Manual correction of the OCR in the <dipl>-layer and manual preparation of the <norm> layer (the original OCR is kept in the <ocr>-layer).
- Part-of-speech tagging and lemmatization with TreeTagger-Batch and TreeTagger for all documents. Please note: Quotation marks can cause errors and need to be masked. Furthermore empty lines will be deleted by the TreeTagger. Fill those lines with a random tag (e.g. <9>) and use the option -sgml while tagging. Lines that include tags will not be tagged and can be deleted afterwards. After merging TreeTagger-output with the MS Excel file, the MS Excel macro SearchAndMerge (Readme) reconstruct the segmentation.
- Automatic creation of <clean> for all documents (Python-Script and Readme).
- Manual alignment of the original <ocr>-layer with <dipl>, <clean> and <norm> (it kept its own segmentation).
- Punctual correction of <norm> and full correction of <persname>. There is a list of unified names now available in the annotation guidelines. In general a lot of small corrections were made all over the corpus.
- Deletion of the following annotation layers in all documents of version 4.1 until 1652: <comp>, <comp_orth>, <prot>. Some of them were replaced by new annotation layers (compare 10.). Deletion of the meta date <preface> in all documents.
- The following annotation layers were added to the new documents of the corpus: <ocr>, <column>, <figure_p>. The following annotation layers were added to all documents of version 6 that were published until 1652: <comp_n>, <comp_n_graph>, <comp_a>, <comp_a_graph>, <infl_fuge>, <adja_uninfl>, <comp_lex>, <comp_amb>, <nlp_morph>.
- The following anotation layers were added to the corpus in the CoNLL format (for all documents of version 6 that were published untl 1652): <deprel>, <morph>, <lemma> (in ANNIS & PAULA <lemma-dep>), <pos> (in ANNIS & PAULA <pos-dep>). They were prepared automatically with the Mate Tools.
- The following annotation layers were added to the corpus in the PTB format (for all documents of version 6 that were published untl 1652): <cat> (in ANNIS & PAULA <cat-const>), <edgelabel> (in ANNIS & PAULA <func>), <lemma> (in ANNIS & PAULA <lemma-const>), <pos> (in ANNIS & PAULA <pos-const>). They were transformed automatically from CoNLL to PTB using the Berkeley Parser.
- All formats (Excel 2013, CoNLL, PTB) were converted to PAULA and ANNIS altogether using Pepper. The following importers were used: SpreadsheetImporter, PTBImporter, CoNNLImporter. Merging with Merger Module. Export with ANNISExporter and PAULAExporter.
Corpus design
| In order to study the development of the scientific language throughout the period of interest, we require a subject domain that is sufficiently well represented in all subperiods. That is why we have selected the domain of herbology (Kräuterkunde). Texts vary somewhat in length since older text is more difficult to annotate. |  |
Annotation layers
The RIDGES corpus is designed as a multi-layer architecutre. Annotation layers can be roughly divided into five kinds:
- Transcription/normalisation
- Linguistic annotations
- Structural annotations
- Content annotation
- Metadata
Transcription/normalisation
| Annotation layer and value(s) | Description |
|---|---|
| dipl independent segmentation annotation value(s):
|
The diplomatic transcription of the word form as found on the manuscript. A Unicode-table with special character is used. |
| clean independent segmentation annotation value(s):
|
Automatic normalization by a Python-Script regarding graphical structures and special characters only (e.g. "ſ" to "s"). For example, a form with a line break like wor=den will be cleaned to worden but not normalized to modern geworden where this would now be the appropriate form. For words including line breaks notice, that if the second word begins with a capital letter, this letter will be normalized to a small letter in the clean layer (e.g. "Gelb- Sucht" to "Gelbsucht"). If all letters of the second word are capital letters, they will remain the same (e.g. "MON- TANUM" to "MONTANUM"). Dipl units containing vowels with macrons are replaced by each potential form of that token, separated by '|' (for example: 'auſzwēdig' to: 'auszwemdig|auszwendig'). For a full overview of the replacements for the clean-tier see the Readme. |
| norm independent segmentation annotation value(s):
|
In this layer the segmentation, graphemics, inflection forms and lexemes are normalized. Graphemics: orthographic normalization according to Duden (e.g. kreutter -> Kräuter); phonology: please notice the sound changes of the Early New High German period, like diphthongization, monophthongization, syncope, apocoke, etc. (e.g. lehret -> lehrt); morphology: in die Nasen -> in die Nase; lexicology: extinct lexical material is normalized according to modern orthography and described in the layer "erlaeuterung" as the case may be (e.g. Vergeſz -> Vergess); word formation: extinct morphemes are normalized - if possible - according to modern orthography (e.g. halben -> halber or stachelecht -> stachelig). Currently there are only some documents in which case was normalized. |
| ocr independent segmentation annotation value(s):
|
This layer was prepared for the new documents of version 7 only. For more information click here. |
Linguistic annotations
| Annotation layer and value(s) | Description |
|---|---|
| pos segmentation based on 'norm' annotation value(s):
|
Autmatic part-of-speech annotation using the STTS tagset for German. |
| lemma segmentation based on 'norm' annotation value(s):
|
Automatic lemmatization by TreeTagger. |
| comment segmentation based on 'dipl' annotation value(s):
|
This is an unsystematic layer. In some cases where the use of modernized orthography is impossible or misleading, a modern semantic equivalent (e.g. Heümonat -> Juli) or an explanation can be given. This layer was originally called "hyperlemma" and was renamed as "erlaeuterung" in ridge-v5. |
| foreign segmentation based on 'dipl' annotation value(s):
|
Non-german text. |
| foreign_trans segmentation based on 'dipl' annotation value(s):
|
Translation from and to German. |
| lang
segmentation based on 'dipl' |
The language of a foreign area is written in (ISO three letter codes according to ISO 639-2). |
| comp segmentation based on 'dipl' annotation value(s):
|
Nominal compound (with nominal head). |
| comp_orth segmentation based on 'dipl' annotation value(s):
|
Annotation of the specific spelling of the annotated compounds (komp): zs: written together, gtr: written separately, bs: hyphenated (one line), lb1: separated by line break (hyphenless), lb2: separated by line break (hyphenated). |
| prot segmentation based on 'dipl' annotation value(s):
|
Assigns prototypes to each compound of the komp-layer : prot1: reliably identifiable as compound, prot2: quite likely a compound und prot3: case of doubt (not assigned in (komp)). |
| comp_amb segmentation based on 'dipl' annotation value(s):
|
Annotation of word orders that are potential compounds. a: adjective-noun sequences in which the adjective is uninflected an which we would produce as compounds in modern German (e. g. das edel geſteine). However, there are a lot of AN sequences including uninflected adjectives that would be attributive adjektvies in modern German as well. These kinds of adjectives were annotated in <adja_uninfl>. gpre: noun-noun sequences that could be compounds as well as nominal phrases with a prenominal attributive genitive. For the annotated instances no clues of context or inflection were available to give information about the grammatical state of these constructions. |
| infl_fuge segmentation based on 'dipl' annotation value(s):
|
Annotation of inflectional elements respectively "Fugenelemente" in word orders that are potential compounds. y: There is an element between word1 and word2 (‚yes‘). n: There is no element between word1 and word2 (‚no‘). yn: Accounts for potential compounds with more than one lexical word parts. In this case there might be an element between two of the words (e.g . Jungkfrawen har). The order of "y" and "n" is strict in this tagset. NA: No information about the existence or non-existence of elements between word1 and word2 can be given, mostly because one of the words is semantically untransparent or one of the words origins from a foreign language (e.g. Latin). |
| comp_lex segmentation based on 'dipl' annotation value(s):
|
lex: lexicalized compounds, that can not be used as a syntagm (anymore), because the addition of the meaning of the word parts is distinct from the compound meaning (affected topics: plant names, geographic terms, some diseases, body parts, animal species, star signs). Given a lexicalized compound has one or more additional parts that do not belong to the lexicalized part, the whole compound does not count as lexicalized, e.g. Eisenkrautsaft or Beifußblumen (vs. Johannisblumen), Blutwassersucht. This works similar for the word "Baum", which can be used to describe a plant more specificly (Kirsche -> Kirschbaum, Eiche -> Eichenbaum) and hence, does not belong to a lexicalized compound. But it also can be part of the lexicalized term, e.g. Schildkraut, Rutelkraut, Wunderbaum). n: This tag is used for all potential compounds that are not being interpreted as lexicalized given the definition above. You can find a list with individual decisions in the more detailed annotation guidelines. |
| comp_n segmentation based on 'dipl' annotation value(s):
|
Additive Tags that give information about the morphologic structure of nominal compounds, e.g. N_N for compounds that consist of two nouns. Suffixes are not incorporated unless suffixation was the last process that was realised (e.g. [[Kindbett]erin]). The tags for each lexical element are separated by underscores. Given the morphologic category can not be identified the palceholder "X" is set. |
| comp_n_graph segmentation based on 'dipl' annotation value(s):
|
The graphemics of the in <comp_n> annotated compounds is described in this layer. sep: seperated; nospace: written together; hyph: hyphenated; lb1: line break (without hyphen); lb2: line break (with hyphen); camel: camel case. Given compounds that consist of more than two lexical parts can be annotated with additive tags, e.g. „nospace_sep“ for Saurampffer waſſer. |
| comp_a segmentation based on 'dipl' annotation value(s):
|
Additive Tags that give information about the morphologic structure of adjective compounds, e.g. A_A for compounds that consist of two adjectives. Suffixes are not incorporated unless suffixation was the last process that was realised. The tags for each lexical element are separated by underscores. farb: the last lexical element contains the root „farb“, e.g. himmelfarben. |
| comp_a_graph segmentation based on 'dipl' annotation value(s):
|
The graphemics of the in <comp_a> annotated compounds is described in this layer. sep: seperated; nospace: written together; hyph: hyphenated; lb1: line break (without hyphen); lb2: line break (with hyphen). |
|
attr_gen
|
Annotation of nominal phrases with genitive attribute post or prenominal. gpre = prenominal genitive, gpost = postnominal genitive. |
|
morph_ellipsis
|
Coordination of compounds and parts of compounds (truncated morphemes and compounds such as: gelb⸗ und Waſſerſucht). |
|
persname
|
Every name of a person to which the author of a particular document refers is annotated. For every instance the name of the person is given in the nominative form. There is a list of unified names now available in the annotation guidelines. |
|
title
|
Every title of books to which the author of a particular document refers is annotated. For every instance the title of the book is given in the nominative form. |
| deprel segmentation based on 'norm' annotation value(s):
|
Dependency annotationen based on the TIGER annotation scheme (prepared with Mate Tools). You can find the tagset in the more detailed annotation guidelines. |
| cat-const segmentation based on 'norm' annotation value(s):
|
Constituency annotation based on the TIGER annotation scheme (automatically transformed out of the dependency parses of the Mate Tools using the Berkeley Parser). You can find the tagset in the more detailed annotation guidelines. |
func
|
Edgelanel annotation based on the TIGER annotation scheme (automatically transformed out of the dependency parses of the Mate Tools using the Berkeley Parser). You can find the tagset in the more detailed annotation guidelines. |
| morph segmentation based on 'norm' annotation value(s):
|
Morphologic annotation with additive tags including information about case| number| gender| tense| comparison| voice (depending on the part of speech). sg/pl: singular/ plural; neut/ masc/ fem: neuter/ masculine/ feminine; nom/ gen/ dat/ acc: nominative/ genitive/ dative/ accusative; 1/ 2/ 3: 1st/ 2nd/ 3rd person; pres/ past: present/ past; pos/ comp/ sup: positiv/ comparative/ superlative; ind/ subj: indicative/ subjunktive; *: placeholder. |
| cat segmentation based on 'norm' annotation value(s):
|
Span annotation, that was created during the conversion process from CoNLL to ANNIS. |
| lemma-dep segmentation based on 'norm' annotation value(s):
|
Lemmatization with Mate Tools. |
| pos-dep segmentation based on 'norm' annotation value(s):
|
Part-of-speech tagging with Mate Tools. |
| pos-const segmentation based on 'norm' annotation value(s):
|
Part-of-speech tagging with Mate Tools and conversion into PTB-Format. |
| adja_uninfl segmentation based on 'norm' annotation value(s):
|
Annotation of uninflected adjectives that are positioned directly in front of a noun. Given there are several uninflected adjectives in front of the same noun only the last one is annotated. |
|
form_disease
|
NA |
|
problem
|
NA |
|
herbname_norm
|
In this layer a systematic herbal name is given. Sometimes it is ambigous - in this case you can find additional information in the "erlaeuterung" or in the "bemerkung_lexik" layer. |
|
herbprep
|
This layer was made for the identification of preparations of herbs. Only those instances are included which are NPs or modifiers with a herb as head. The name is given in the nominative singular form and normalized according to modern orthography. Whitespaces are replaced by underscores. Compounds are always written together, regardless of their compound spelling in the facsimile. Everything is written in lower case letters (e.g. safft des weremuts -> saft_des_wermuts. |
|
form_prep
|
In this layer preparations with herbs are described syntactically or morphologically. kompNN = NN compounds which are written together or hyphenated; kompNNgetrennt = nouns following each other which could be a compound (written seperatlely); phraseVon = preparations with herbs containing a von-PP (e.g. safft von weremut); phraseGen = preparation with herbs containing a genitive attribute (e.g. safft des weremuts. |
|
noun_nom
|
In this layer all nouns which are included in the text are given, namely in the first occurring spelling in the nominative singular form. If the first occurring form of "Saft" is safft, all further incidences of "Saft" are given as safft. Everything is written in lower case letters. The purpose of this layer is to investigate the variation of noun spelling within one text. |
|
form_noun
|
In this layer all nouns are were morphologically annotated. kompNN = NN compound, written together or hyphenated; kompNNgetrennt = all sequences of two nouns which could be compounds, but are written separately; kompNEN = NE-N compound, written together or hyphenated; kompNENgetrennt = all sequences of NE and N which could be compounds, but are written separately; kompNNNgetrennt = all sequences of three nouns which could be compounds, but are written separately; kompAN = AN compounds; kompVN = VN compounds; ; derivat = derivates; nom= implicite nominalisation (conversion, ablaut, syntactic, nominalisation); gri/lat/ara = clear Greek/Latin/Arabic nouns, already in the German language integrated foreign material is treated like native words; lex = lexicalized herb names which were originally morphologically complex, e.g. Beifuß, Wermut, Stabwurz, and tausend guldin for "Tausendguldenkraut". |
|
comment_lex
|
This is an unsystematic layer for comments and questions about lexis. |
|
clause_type
|
Annotation of clause types. No hierarchical annotation. For nested sentences only the highest clause is annotated. In the layer "bemerkungen_syntax" you can find notes about the nestings. rs = clear relative clauses, both "w-relative clauses" and "d-relative clauses"; padv = clauses which are introduced by a pronominal adverb; rsx = relative clauses without main clause (this often occurs in headlines); rsdem = ambiguous cases: relative clause or demonstrative clause; padvpart = clauses with pronominal adverb and participle; dem = demonstrative clauses (all clauses with a demonstrative pronoun as subejct); part = participles that behave similarly like relative clauses. |
|
position_rel
|
Position of the relative clause within the main clause. vor = preposed; nach = postposed; int = embedded. |
|
position_referent
|
Position of the relative clause relative to the reference category. adja-v = adjacently preposed; adja-n = adjacently postposed; dist = distant; na = not applicable. |
|
form_referent
|
Form of the reference category of the relative clause. np = non-pronominal NP; d-pron = der, die, das, dieser, etc.; p-pron = personal pronoun; null = for free relative or asyndetic relative clauses with a covert correlate in the main clause). |
|
position_verb_rel
|
Verb position within the relative clause. v2 = verb second; ve = verb end; venf = verb end with occupied postfield. |
|
form_relpron
|
Form of the category which introduces the relative clause.d-pron = all d-pronouns; w-pron = wer, welch-; w-phras = e.g. welch frau |
|
mod_referent
|
relsatz = Annotated on pronouns, NPs or clauses, if modified by a relative clause. Not applicable for free relative clauses. The whole reference category is annotated as span.d-pron/m-padv, m-part, np = NA. |
|
position_verb
|
Verbposition.V2: Verb second position at a given subordinated clause with subordinated conjunction; analyzed as a token feature at occurrences of pos=KOUS. Vletzt: Verb final position at a given subordinated clause with subordinated conjunction; analyzed as a token feature at occurrences of pos=KOUS. V?: Unclear verb position at a given subordinated clause with subordinated conjunction; analyzed as a token feature at occurrences of pos=KOUS.V1: Verb first position at a given subordinated clause with subordinated conjunction; analyzed as a token feature at occurrences of pos=KOUS. |
|
subclause_type
|
Type of subordinating clause. Adverbial: Adverbial function of a given subordinated clause with subordinated conjunction; analyzed as a token feature at occurrences of pos=KOUS. Attribut: Attributive function of a given subordinated clause with subordinated conjunction; analyzed as a token feature at occurrences of pos=KOUS. Komplement: Complement function of a given subordinated clause with subordinated conjunction; analyzed as a token feature at occurrences of pos=KOUS. |
|
KOUS_sem
|
KOUS_Semantik. additiv: Additive semantics of subordinated conjunction; analyzed at occurrences of pos=KOUS. final Final semantics of subordinated conjunction; analyzed at occurrences of pos=KOUS. k.a.: Non analyzable semantics of subordinated conjunction, due to complement status of subordinated clause; analyzed at occurrences of pos=KOUS.kausal Causal semantics of subordinated conjunction; analyzed at occurrences of pos=KOUS. konditional: Conditional semantics of subordinated conjunction; analyzed at occurrences of pos=KOUS. konsekutiv: Consecutive semantics of subordinated conjunction; analyzed at occurrences of pos=KOUS. konzessiv: Concessive semantics of subordinated conjunction; analyzed at occurrences of pos=KOUS. modal: Modal semantics of subordinated conjunction; analyzed at occurrences of pos=KOUS. temporal: Temporal semantics of subordinated conjunction; analyzed at occurrences of pos=KOUS; 0: NA. |
|
ppk_e1 - ppk_e3
|
Prepositional constructions (prepositional attributive constructions or attributive adverbial phrases) are annotated. ppk: normal prepositional construction ppk_e2: normal ppk inside a ppk in the layer <ppk_e1> ppk_e3: normal ppk inside a ppk in the layer <ppk_e2> zwf: case of doubt ppk_rek: recursive (nested) ppk attr_X: attributes, that refer to an element inside a ppk without linking to that inside a syntactic sequence. X is a placeholder for the respective referent. |
|
sentence_end
|
Sentence endings are annotated. You can find the detailed annotation guidelines here. |
Structural annotations
| Annotation layer and value(s) | Description |
|---|---|
|
lb
|
Linebreak. |
|
pb
|
Pagebreak. |
|
pb_n
|
The number of the page (if marked explicitly). |
|
pb_ana
|
Revision of the pagebreak (e.g. in case of apparently incorrect page numbers). |
|
unclear
|
Unreadable or otherwise unclear text. |
|
atLeast
|
Minimum presumed length of unclear text in characters. |
|
atMost
|
Maximum presumed length of unclear text in characters. |
|
interpretation
|
Suggestions for unreadable or unclear text. |
|
figure
|
A graphic or table embedded in the original document. |
|
figure_rend
|
Description of the rendering of a figure. |
| figure_p segmentation based on 'dipl' annotation value(s):
|
Annotation of the original page in the facsimile on which a figure is printed. |
| column segmentation based on 'dipl' annotation value(s):
|
Annotation of all dipl-units that belong to one column on a page. This annotation is only annotated when there is more than one column per page. l: left; r: right. |
|
hi
|
Highlighted area. |
|
script
|
Annotation of change of font. |
|
hi_rend
|
Description of the rendering of the highlighted area. |
|
head
|
A heading. |
|
note
|
A note in the original document (e.g. footnotes, margins). |
|
ref
|
Reference to a footnote. |
|
ref_target
|
ID of the footnote being referred to. |
|
ref_type
|
Type of reference (e.g. a TEI "noteAnchor"). |
|
quote
|
dipl-tokens that are part of a quote are annotated with "yes". The default value is "no". |
|
item
|
Item in a list. |
Content annotations
These annotations were developed by our students to annotate spans of tokens with properties of special interest.
| Annotation layer and value(s) | Description |
|---|---|
|
definition
|
|
|
author_ref
|
References made by authors to themselves. Values indicate the grammatical type of the reference, e.g. "pron1pl" for first person plural pronoun. |
|
reader_ref
|
References made by authors to the reader. Values indicate the grammatical type of the reference, e.g. "pron2sg" for second person singular pronoun. |
|
plant
|
Naming of a plant |
|
property
|
Description of properties like appearance, smell, etc. |
|
name
|
A proper name (annotated only in some documents). |
|
name_type
|
The type of proper name (e.g. "person", "herb"). |
|
reference
|
This unsystematic layer is for referencing interpretations of all kind. |
Metadata
These annotations are loosely based on the TEI P5 guidelines. Furthermore you can find the complete corpus meta data in TEI p5 here: HANDLE ID. All meta data are annotated for each document.
| Annotation layer and value(s) | Description |
|---|---|
|
author
|
Name of the author (if known). |
|
bibl
|
Full bibliographical entry for the source including the page numbers annotated in the corpus. |
|
date
|
Date of publication, usually just the year (e.g. "1722"). |
|
publisher
|
Publisher of the document (if known). |
|
place
|
Publication place of the document. |
|
title
|
Title of the work the document was extracted from. |
|
translator
|
Translator of the text, if existing. |
|
trans_from
|
Language from which the text was translated. |
|
editor
|
Editor of the text, if known.. |
|
version
|
Version in which the specific document was added to the corpus. |
|
edition_first
|
Erstauflage: first edition of the text; Nichterstauflage: not the first edition of the text. |
|
issue
|
Volume of the text, if known. |
|
maintopic
|
science: the text is about scientific topics; non-science: the text is about everyday topics. |
|
topic
|
One or more topics per text are given. Additive value in alphabetical order of the abbreviations. Al: alchemy, As: astronomy, B: botany, G: gardening, K: kitchen, L: linguistics, M: medicine, R: religion. Example values: "B", "BM" oder "BKM". |
|
register
|
Register of the text: Herbology. |
|
lingualism
|
mehrsprachig: the text is multilingual, which means that there are whole paragraphs written in another language than German (single translations of specialist terms do not count); einsprachig: the text is monolingual. |
|
orig_date
|
If a text is categorized as "Nichterstauflage" in "auflage", the original date of publication is given here (if known). |
|
orig_place
|
If a text is categorized as "Nichterstauflage" in "auflage", the original place of publication is given here (if known). |
|
repository
|
URL to the repository where you can find the facsimile of the text. |
|
lang_type
|
The language type is given. fnhd: Early New High German, nhd: New High German |
|
lang_area
|
The language area is given. md: Middle German, obd: High German. If a text is a later and more standardised one, the value "NA" is given. |
|
text_type
|
Declaration of the general text composition. Prosa: the text is prosaic, Poesie: the text is poetic; gemischt: the text is partly poetic and partly prosaic. |
|
lyric_type
|
If in "textgestaltung" the values "Poesie" or "gemischt" are given, you can find here the specific poetic elements that are used. Endreim: end rhyme; Metrik: metrics |
|
wormwood
|
ja: there is a paragraph about the topic "Wermut" in this document; nein: there is no paragraph about the topic "Wermut" in this document. |
|
herb_sorting
|
ja: the document is a herbal monography collection, which means that different herbs are described in an ordered selection; nein: the document is no herbal monography collection |