Sentence Parser¶

The Sentence Parser is one of the two features that make LinguaMorpha's Latin module stand alone on consumer iOS. Paste any Latin sentence — Caesar's opening, a Vulgate verse, a line from Cicero, your homework — and the app shows you a full parse of every word.

It lives in Self Study → Identify a Form, where the top toggle switches between Word mode (lookup a single inflected form) and Sentence mode (parse a whole sentence).

A worked example¶

Paste this into the sentence box:

Gallia est omnis dīvīsa in partēs trēs.

(Gaul is, as a whole, divided into three parts.)

Tap Parse. The app renders six tappable chips:

Chip	Parse
Gallia	proper noun, fem. — Gaul. Nom. sg.
est	sum, esse — to be. 3sg. pres. ind. act.
omnis	omnis, omne — all, every. Nom. sg. fem. (in agreement with Gallia).
dīvīsa	dīvidō, dīvidere, dīvīsī, dīvīsum — to divide. PPP, nom. sg. fem. (predicate participle with est → periphrastic perfect passive).
in	preposition, takes acc. or abl. (here acc. — partēs is accusative).
partēs	pars, partis (f., 3rd-decl i-stem) — part. Acc. pl. fem.
trēs	numeral trēs (declinable). Acc. pl. (agreeing with partēs).

Tap any chip to jump to the full lemma view — the verb's conjugation table, the noun's declension paradigm, the adjective's full agreement table, and so on.

How ambiguity is handled¶

Latin words frequently have multiple parses. Vēris could be:

genitive singular of vēr (spring) — second-declension neuter
second-person singular future indicative active of vereor (to fear / revere) — semi-deponent

The chip shows every parse, stacked. Each parse is ranked by pedagogical likelihood — the one most likely in the surrounding context is highlighted; the others are dimmed but selectable. The ranking is heuristic, not semantically intelligent — it favours forms whose lemma is in the high-frequency tier of the catalogue, and forms whose gender / number matches an adjacent agreement target. It's right most of the time but you should treat it as a hint, not a verdict.

For forms with three or more parses, the chip shows the top two by default and an "expand" affordance reveals the rest.

What the parser handles¶

Enclitics — -que, -ve, -ne are split off and parsed independently. Filiusque parses as filius + -que.
Capitalisation — sentence-initial capitals and proper-noun capitals are folded for matching; the original capitalisation is preserved in the display.
Trailing punctuation — . , ; : ? ! are stripped before lookup. Quotation marks (" and ') and parentheses are also handled.
Macron-folded matching — paste Gallia est omnis divisa without macrons; the parser finds dīvīsa anyway. The display preserves the macrons from the catalogue, not from the input.
Compound forms — the periphrastic perfect passive (amātus est), pluperfect passive (amātus erat), and future perfect passive (amātus erit) are recognised when the participle and the form of esse are adjacent.
Unknown words — if a form isn't in the catalogue, the chip shows the word verbatim with a "no parse" indicator. The catalogue has around 750 lemmas in v1.0; rarer vocabulary will hit "no parse" until the catalogue grows.

What it doesn't handle:

Word order or syntax — the parser is morphology-only. It tells you what each word is; it doesn't tell you what each word is doing in the sentence (subject, object, indirect object, agent of passive, etc.). Syntax-aware parsing is a v2.x bet, not v1.x.
Elision — Latin verse elides vowels at word boundaries (atque ego → atqu' ego). The parser does not reconstruct elided forms; paste the unelided text if you want a parse.
Reading-order disambiguation — if a Latin sentence has two valid readings (a real Caesar problem), the parser shows all the parses for each word and lets you choose. It does not pick a reading for you.

Settings¶

Show alternative parses by default — toggle in Settings. When on, the chip always shows all parses; when off (the default), only the top-ranked parse is shown unless you tap to expand.
Show case labels in the chip — toggle. When on, the chip prefixes each parse with the abbreviated case label (nom., gen., dat., …) for fast scanning.

When to use it¶

Reading any Latin text — Caesar, Cicero, Vergil, the Vulgate, your textbook's reading passages. Paste the sentence, get the morphology, then work out the syntax yourself. The parser is the dictionary-and-paradigm-table you don't have to flip pages in.
Cross-checking your own translation — if you've translated a sentence and you're not sure you got a form right, paste it and see what the parser thinks.
Sight-reading practice — paste a sentence cold, see how many words you can identify without the parser, then check yourself with the parser.
Studying ambiguous forms — the parser is also the fastest way to see how often a given form (say populī) is ambiguous in real texts. Paste the sentence, see the multiple parses.

The parser is not a substitute for learning Latin. It's a faster reference than flipping through a textbook's appendix.

Roadmap¶

v1.1 — Expanded catalogue and a few more high-frequency lemmas; better ranking of ambiguous parses; "common phrase" recognition (e.g. eō ipsō, nē quid).
v1.5 — Etymology bridge: hover any parsed Latin word to see its Romance descendants (Spanish partes, French parties, Italian parti, Portuguese partes).
v2.x — Syntax-aware annotations: subject / object / agent labels on chips, agreement arrows between agreeing words. Big bet; not committed.