Inflections

The inflection records how a lemma is realised in this token: case, number, tense, and so on. It is what the Viewer prints under each word — viharaḍi / vi + hṛ, pres. sg. 3rd.

The fields follow the part of speech

You do not choose which fields to fill in. The token inspector shows only those meaningful for the lemma's part of speech and subpart, and hides the rest.

Part of speech	Fields shown
noun	Nominal Gender · Case · Grammatical Number
adj., number	Case · Nominal Gender · Grammatical Number
pron. — pers., refl.	Case · Grammatical Number
pron. — other subparts	Case · Nominal Gender · Grammatical Number
v. — Finite	Grammatical Number · Verbal Person · Verbal Voice · Verbal Tense · Verbal Mood
v. — Derivative	Verbal Secondary Conjugation
ind., ptc., adv., adp., and other flat terms	none

Indeclinables have no inflection section at all — the heading disappears with the fields.

A verb with no inflection fields is missing its subpart

A lemma whose part of speech is v. but which has no subpart chosen shows no inflection fields, because Finite and Derivative take completely different ones. Set the subpart on the lemma and the fields appear.

Values

Each field is a dropdown from a fixed vocabulary, and each can be cleared back to —. A partial analysis is perfectly acceptable — record what the form actually tells you and leave the rest empty.

Verbal voice offers active (P.) and middle (Ā.).

Changing the lemma changes the fields

Repointing a token at a lemma with a different part of speech re-computes which fields apply. Values belonging to fields that no longer apply are not saved — a case marking does not survive a switch from noun to verb. Check the inflection after changing a lemma.

Cognates

Cognates are properties of the lemma, not the token: a language and a form, most often the Sanskrit equivalent. A lemma may carry several. Because they sit on the lemma, every attestation inherits them, which is what lets the Viewer show the Sanskrit alongside any occurrence of the word.