Part of Speech and Morphology Search Card
There are two ways to explore the corpus in this card:
- Filter: pick a work, part of speech with grammatical features, and/or a partial match for a word, and MOPSOS will show matching forms.
- Custom queries: for advanced users, custom SQL queries can be written.
1. Filter
View words that appear in the texts (Form) and their dictionary form (Lemma). You can filter by work, part of speech, or a sequence of letters. Results are paginated.
#abc searches the start of the word, abc# the end, and #abc# finds an exact sequence match. ab*c searches for words that contain ab followed by any number of letters then c. You can also use regular expressions to query forms and lemmas.
Read-only SQL over table morphology. Reserved-word columns are quoted, e.g. WHERE "case" = 'g'. Results are not paginated.
Ready: write SQL and press Enter.
Part of Speech and Morphology Distributions Card
In this card, aggregate information about parts of speech in the corpus can be displayed:
- Distribution Charts: count words by any dimension (book, lemma, case, etc.), optionally broken down by a second dimension, under any filter. For example: which books have the most aorist optatives, or how noun number and case combine.
1. Distribution Charts
Group words by morphosyntactic and textual attributes and visualize their distributions. For example, if you set part of speech to Verb, mood to Optative, tense to Aorist, and count by Book / work, the bar chart will show how many instances there are of aorist verbs in the optative within the works in the current corpus.
Charts the distribution of just these lemmata — one alone, or several together. Set Count by (or Break down by) to Lemma to compare them side by side, or to Book / work, Case, etc. to see where and how they occur.
Greek and Beta Code match the corpus lemma list directly (accent-insensitive). The network shows the words that actually share a sentence with the chosen lemma, counted live from the corpus.
1. Nominal compounds
Nominal compounds panel
Filter the analyzed compounds by first member, second member, member category, stem subcategory (s-stem, thematic, and so on), or attesting work. The matching compounds are listed first (each with its segmentation and, for both slots, the exact compound member alongside its lemma and category); the category-pairing figure and its table follow, then the commonest members and the metrical localization.
- Members: search a substring of the member as it surfaces in the compound — the string before the
+for the first member, after it for the second (Greek with or without accents, or Beta Code; or click to browse a member). The+anchors the morphological boundary and#the word edge: in the first-member boxο+= first members ending in ‑ο right before the boundary (the linking ‑ο) and#ἀγα= beginning the word; in the second-member box+μων= second members beginning with μων right after the boundary andμων#= ending the word. Because the search runs on just that side of the+, an anchored search is effectively a stem-class search. Alternatively pick a member category or subcategory. The two slots stay mutually compatible: choosing a first member (or its category or subcategory) narrows the second-member lists to the values that actually combine with it, and vice versa. - Attested in: any corpus work (Iliad, Odyssey, Theogony, Works and Days) — a compound counts as attested in a work when its lemma occurs among that work's tokens (lemma matching across the database).
- Lookup: type Greek (accents optional), Beta Code, or English; substrings match anywhere, and
#anchors the word edge —#ἀγα= starts with,μων#/mwn#= ends in ‑μων (Beta Code writes long o as w and long e as h, soths#= ‑της). Choosing one compound shows its members (exact member and lemma), every metrical position it takes in the verse, every occurrence in the corpus with its line text, and a button that shows all of its lines scanned word by word.
Matching compounds
Compound member categories
Members of the matching compounds
The commonest first and second members among the compounds matching the filters above. Each lexeme's bar is split by allomorph in descending order of attestation (hover a segment for its exact share). Click a member to open a chart just below with each of its allomorphs counted separately.
Metrical localization
Show generated SQL
Infinitive Search Card
This card searches the allomorphs of the infinitives.
- Allomorph filters: restrict by tense, voice, conjugation class (athematic, thematic, contract, sigmatic), or ending type. The ending type can be chosen on its own (each option names its class) or after a conjugation, which narrows the list to that class's endings. We only have active and passive infinitives tagged; medial and mediopassive infinitives are currently being processed. Each result row carries the infinitive's tense, voice, conjugation class, ending type, lemma, and the metrical shape of its stem and whole form in situ. An example search like ζευγνύμεναι would return present-tense, athematic conjugation, ending ‑μεναι, lemma ζεύγνυμι, and stem shape HL.
- Metrical filters: restrict by the metrical shape of the whole form or of its stem (H = heavy/long, L = light/short, one letter per syllable); type a shape and the menu narrows to the attested ones, or click to browse.
- Word searches: Form contains and Lemma contains find substrings, accent-insensitively;
#abcsearches the start,abc#the end, and you can toggle to regular expression search by clicking the box. Lemma matches offers only the lemmata with attested infinitives as you type. - Custom queries: for advanced users, custom SQL queries can be written.
1. Filter
Find infinitives by scope, grammar (tense, voice, conjugation, ending morpheme), metrical shape, lemma, or form. Choose any values and Apply; results are paginated.
#abc searches the start of the word, abc# the end, and #abc# finds an exact sequence match. You can also use regular expressions to query forms and lemmas.
Read-only SQL over table morphology. Reserved-word columns are quoted, e.g. WHERE "case" = 'g'. Results are not paginated.
Ready: write SQL and press Enter.
2. Visualizations
Build descriptive figures of infinitive allomorphy straight from the corpus. Pick a figure, optionally restrict to one work and set a minimum token count for the per-row breakdowns, then Generate. Every figure can be enlarged and downloaded as PNG or SVG.
Optionally restrict which infinitives feed the figure, exactly as in the Filter card above. These narrow the slice before it is charted; leave them at (any) to chart everything.
Chart the result of your own read-only SQL over table morphology instead of a preset figure. Two result shapes are recognised:
- Allomorph stack (the paper's figures): return a label column plus
conjugation,tense,morpheme, optionally a count column namedc,n,countorfreq. Each row is bucketed into its ending allomorph by the same (class, tense, number) map the presets use, and bars are grouped by the first non-count column. Example:SELECT work AS label, conjugation, tense, morpheme FROM morphology WHERE pos='v' AND mood='n' AND voice='p'. - Generic bars: return exactly two columns
(category, value)for a bar chart, or three(group, series, value)for a stacked bar. Reserved-word columns must be quoted, e.g."case".
Ready: write SQL and press Run & chart.
Morphotactics tab
Morphotactics is the study of how morphemes and morphological features are ordered and combined. Each panel explores a different aspect of the corpus:
- Nominal compounds: filter the analyzed compounds by a specific first or second member (adaptive Greek or Beta Code input, with a browsable list of attested members) or by the members’ syntactic categories or stem subcategories (s-stem, thematic, and so on, wherever the analysis records one), and by attesting work. The panel shows how the member categories pair up, the commonest first and second members among the matches, and where the matching compounds are localized in the hexameter (from the merged metrical record); the lookup gives one compound’s members, every metrical pattern it shows, every occurrence in the corpus with its line text, and a button that scans all of its lines word by word.
- Infinitive forms: tense/voice combinations attested among infinitives, optionally restricted to a work; includes a lookup for an individual verb's attested infinitive forms.
- Word-class sequencing: a transition matrix of part of speech to next part of speech for adjacent tokens within a sentence.
- Feature co-occurrence: the co-occurrence of one morphosyntactic component with another inside the same word, optionally restricted to one part of speech.
- Paradigm slots: how often each value of a feature is filled for a chosen part of speech; only the features that actually apply to that part of speech are offered.
Only the features that actually apply to the chosen part of speech are offered.
Result
Pick a view above and click “Render view”.
Show generated SQL
Stylometry tab
Each unit you cluster (a work, an author, a part-of-speech class…) becomes a "document". Its features are the frequencies of the tokens it contains, turned into a profile and compared with a distance measure; units that come out close together are grouped into the same cluster.
- Cluster by sets what each point represents. For part of speech, number, and case you can pick a specific value (e.g. cluster works by how they use verbs) or cluster the categories themselves.
- Limit to restricts the corpus before features are built: by work, by author, or by grammar (pick a part of speech, then add only the features that apply to it).
- Advanced features / clustering options expose the feature model, distance metric, method, and number of clusters. Sensible defaults are used otherwise.
1. Configure features
Pick what each point represents and (optionally) restrict the data. Everything else has sensible defaults under Advanced features.
Choose a part of speech first; then only the features that apply to it can be added, so compatible attributes combine (e.g. tense + number on a verb) while incompatible ones (e.g. person + case) never appear together.
2. Clustering strategy
Sensible defaults are applied. Open Advanced clustering options to change the method or the number of clusters.