Yezur Wiki:Ambiguous glosses/Part two

This is the evidence behind part two of Yezur Wiki:Ambiguous glosses — the glosses that are a single English word which English itself splits. The parent page said only that a row was "unqualified"; that is not enough to rule from, so every entry behind every pending row is set out here with what its own definition says, not merely its gloss field.

Each gloss carries one of two verdicts:

Answerable from the entries

The definitions already settle it and only the short gloss field is out of step. No ruling is wanted — the proposed new gloss text is given, and it can simply be applied.

Needs a ruling

Every entry really is bare, or the distinction was never decided. The question is stated once, with its options; no answer is invented.

Of the 459 entries behind these rows, 285 carry a definition saying more than the bare gloss. That is why so much of this list turns out not to need the operator at all.

Two different words, spelled alike

fly

Two unrelated English words: the insect, and the act of moving through the air. Part of speech separates all fourteen entries cleanly.

Language	Word	Current gloss	What the entry's own definition says
Proto-Herta-Olaic	*ʃeiɹhə	fly	verb — "fly, move through the air"; by extension "move fast"
Ancient Hertic	ςερχισ	fly	verb — "fly"
Gaillean	cehirei	fly	verb — "fly"
Middle Airananwe	ceyeree	fly	verb — "fly"
Airananwe	cheryé	fly	verb — "fly"
Proto-Njeshan	*gjuran	fly	verb — "fly"
Proto-Andusian	*kpuŋke	fly	verb — "fly"; "take wing"; by extension "leap, spring"; figurative "to pass swiftly, as time"
Proto-North-Andusian	*kpuŋke	fly	verb — "fly"; "take wing"; by extension "leap, spring"
Proto-South-Andusian	*kuŋke	fly	verb — "fly"; "take wing"; by extension "leap, spring"
Proto-Unterian	*puŋke	fly	verb — "fly"; "take wing"; by extension "leap, spring"
Besoh	kiunke	fly	verb — "to fly; take wing"
Gaillean	connu	fly	noun — "fly"
Middle Airananwe	conge	fly	noun — "fly"
Airananwe	chong	fly	noun — "fly"

Answerable from the entries. Eleven verbs take fly (move through the air) — the wording *ʃeiɹhə already uses: *ʃeiɹhə, ςερχισ, cehirei, ceyeree, cheryé, *gjuran, *kpuŋke, *kpuŋke, *kuŋke, *puŋke, kiunke. Three nouns take fly (insect): connu, conge, chong. Nothing in the dictionary offers a competing nominal reading such as "a flight" or "a flier", and the three languages that carry the noun each carry a separate general word for the class of creature — kennie, knyi, gny "insect" — so the noun is the particular creature. The Descendants list on *ʃeiɹhə derives both chains from the one verb, which is consistent with this and does not disturb it.

stick

Two unrelated English words: the piece of wood, and the verb "to stick" (adhere, thrust). All ten entries are nouns, and six spell the wooden sense out.

Language	Word	Current gloss	What the entry's own definition says
Proto-Herta-Olaic	*dβexa	stick	noun — "stick; a piece of wood, a staff or rod"
Ancient Hertic	βεξι	stick	noun — "stick"
Gaillean	dveza	stick	noun — "stick"
Middle Airananwe	dueesye	stick	noun — "stick"
Airananwe	duése	stick	noun — "stick"
Proto-Andusian	*mutkɔ	stick	noun — "stick, branch"; "staff"; by extension "club, tool-handle"
Proto-North-Andusian	*mutko	stick	noun — "stick, branch"; "staff"; by extension "club, tool-handle"
Proto-South-Andusian	*mutkɔ	stick	noun — "stick, branch"; "staff"; by extension "club, tool-handle"
Proto-Unterian	*mukɔ	stick	noun — "stick, branch"; "staff"; by extension "club, tool-handle"
Besoh	mukko	stick	noun — "stick, branch"; "staff"; by extension "club, tool-handle"

Answerable from the entries. All ten take stick (piece of wood): *dβexa, βεξι, dveza, dueesye, duése, *mutkɔ, *mutko, *mutkɔ, *mukɔ, mukko. No verb "stick", and no verb "adhere", is attested in any language; the Andusian five define the word by its material referents (branch, staff, club, tool-handle), and *dβexa says "a piece of wood" outright, comparing itself to *dwɑl "tree". Each branch also keeps "wood" as a separate lemma (*sim, trolliquu), so the noun is the object, not the material.

right

Two unrelated English words, both adjectives: the hand and side opposed to left, and "correct". Part of speech settles nothing here. The pages be, *be and bë carry several unrelated lemmas under one spelling, split by numbered Etymology sections; only the adjective section is quoted below.

Language	Word	Current gloss	What the entry's own definition says
Proto-Herta-Olaic	*ətʒɛnz	right (hand)	adjective — "right; the strong side"; by extension "correct and proper"
Ancient Hertic	ατιηνε	right (hand)	adjective — "right"
Gaillean	ezhenna	right (hand)	adjective — "right"
Middle Airananwe	ezyenga	right (hand)	adjective — "right"
Airananwe	ejnay	right	adjective — "right" (etymology undetermined)
Proto-Andusian	*ᵐbe	right (hand)	adjective — "right(-hand)"; by extension "the strong side"; "correct" (third sense, unlabelled)
Proto-North-Andusian	*ᵐbe	right (hand)	adjective — "right(-hand)"; by extension "the strong side"
Proto-South-Andusian	*be	right (hand)	adjective — "right(-hand)"; by extension "the strong side"
Proto-Unterian	*me	right (hand)	adjective — "right(-hand)"; by extension "the strong side"
Besoh	be	right (hand)	adjective — "right(-hand)"; by extension "the strong side"
Besoh	bë	right (hand)	adjective — "right(-hand); the strong side"

Answerable from the entries. Ten of the eleven are already qualified and need no change. Only ejnay is bare, and it takes right (hand). The ground is Airananwe's own vocabulary rather than the entry, which carries {{etym-missing}}: the language separately records thalcey "correct" (from tyalce, tvalla, *tβɑlj) and tibhiey for the opposite hand, so the sense "correct" is occupied and the sense opposed to "left" is not. Every one of the ten languages in the table likewise carries a distinct adjective "correct" — *tβɑlj, τφαιε, tvalla, tyalce, thalcey, *moᵑg, *moᵑg, *moŋ, *moŋu, mon — and a distinct adjective for the other hand. Note that *ətʒɛnz lists no Airananwe descendant and ezyenga no descendants at all, so ejnay's membership of that set is inferred from form and is not written anywhere; the ruling rests on the Airananwe evidence above.

Two residues, neither a question about sense. be and bë are a Besoh doublet — bë derives from the same Proto-South-Andusian *be "with lengthening of the vowel" — and will read identically in a translations box. And *ᵐbe records "correct" as a full third sense with no by extension label, unlike *ətʒɛnz, although Proto-Andusian also has *moᵑg "correct" and no descendant of *ᵐbe keeps the sense.

drink

Two English words: the verb, and the noun "a drink" (a beverage). All eleven entries are verbs.

Language	Word	Current gloss	What the entry's own definition says
Proto-Herta-Olaic	*ʎiβtʲə	drink	verb — "drink; swallow liquid, take in water"
Ancient Hertic	λιτκισ	drink	verb — "drink"
Gaillean	libtei	drink	verb — "drink"
Middle Airananwe	liptee	drink	verb — "drink"
Airananwe	litté	drink	verb — "drink"
Proto-Njeshan	*bakrib	drink	verb — "drink"
Proto-Andusian	*mobvi	drink	verb — "drink"; "swallow liquid"; by extension "absorb, soak up"; figurative "to drink in, listen or watch raptly"
Proto-North-Andusian	*mobvi	drink	verb — "drink"; "swallow liquid"; by extension "absorb, soak up"
Proto-South-Andusian	*mobvi	drink	verb — "drink"; "swallow liquid"; by extension "absorb, soak up"
Proto-Unterian	*movi	drink	verb — "drink"; "swallow liquid"; by extension "absorb, soak up"
Besoh	mobvi	drink	verb — "drink"; "swallow liquid"; by extension "absorb, soak up"

Answerable from the entries. All eleven take drink (swallow liquid) — the wording both *ʎiβtʲə and the whole Andusian set already use: *ʎiβtʲə, λιτκισ, libtei, liptee, litté, *bakrib, *mobvi, *mobvi, *mobvi, *movi, mobvi. No noun "a drink" or "beverage" is attested in any language, so no entry is in doubt.

dream

Two English words: the noun (a vision in sleep) and the verb. Part of speech splits the eight entries five to three, and the five nouns define themselves.

Language	Word	Current gloss	What the entry's own definition says
Proto-Andusian	*ɲoro	dream	noun — "dream"; "sleep-vision"; by extension "omen"
Proto-North-Andusian	*ɲoro	dream	noun — "dream"; "sleep-vision"; by extension "omen"
Proto-South-Andusian	*ɲoro	dream	noun — "dream"; "vision"
Proto-Unterian	*noro	dream	noun — "dream"; "omen"
Besoh	ñoro	dream	noun — "dream"; "vision"
Gaillean	mbireiei	dream	verb — "dream"
Middle Airananwe	mbireyee	dream	verb — "dream"
Airananwe	biriné	dream	verb — "dream"

Answerable from the entries. The five nouns take dream (sleep-vision) — the Andusian entries' own word: *ɲoro, *ɲoro, *ɲoro, *noro, ñoro. The three verbs take dream (see visions in sleep): mbireiei, mbireyee, biriné. The nouns' extended senses "omen" and "vision" both presuppose the sleeping sense, and the dictionary records no aspiration sense anywhere, in any part of speech. The three verb lines are themselves bare and no noun "dream" exists in the Herta-Olaic branch to fix them internally, so they rest on that absence; the nouns do not.

play

Two English words: the verb, and the noun "a play" (a staged work). All ten entries are verbs.

Language	Word	Current gloss	What the entry's own definition says
Proto-Herta-Olaic	*ɑkhjo:ən	play	verb — "play, sport, do without purpose"; by extension "dance and perform"
Ancient Hertic	αιομισ	play	verb — "play"
Gaillean	aqownei	play	verb — "play"
Middle Airananwe	aqounee	play	verb — "play"
Airananwe	aqoené	play	verb — "play"
Proto-Andusian	*gidz	play	verb — "play"; "frolic"; by extension "dance, make music"
Proto-North-Andusian	*gidz	play	verb — "play"; "frolic"; by extension "dance, make music"
Proto-South-Andusian	*gidz	play	verb — "play"; "frolic"; by extension "dance, make music"
Proto-Unterian	*gizi	play	verb — "play"; "frolic"; by extension "dance, make music"
Besoh	gij	play	verb — "play"; "frolic"; by extension "dance, make music"

Answerable from the entries. All ten take play (amuse oneself): *ɑkhjo:ən, αιομισ, aqownei, aqounee, aqoené, *gidz, *gidz, *gidz, *gizi, gij. Both proto-forms gloss the first sense as purposeless sport or frolic, and no noun "a play" is attested anywhere. The music and dancing sense that six entries record is marked by extension and sits inside the verb; it is not the other English word, and the branches keep separate lemmas for the neighbouring notions in any case — *gida "dance" (with reflexes in Proto-South-Andusian, Proto-Unterian and Besoh) and, in Proto-North-Andusian, *gira "play; contest, game", itself derived from *gida.

Mass against count

The shape of the problem

Sixteen glosses, 165 entries, every one of them a noun. Sixty-eight entries carry a single definition line that repeats the gloss word and adds nothing. They are not scattered: Gaillean (14), Middle Airananwe (14), Airananwe (18), Ancient Hertic (15) and Proto-Njeshan (7) account for all of them, and there is not one among the Proto-Herta-Olaic or the six Andusian stages, whose lines always say more than the headword — hoĥ, one and five Proto-Herta-Olaic entries give only one line, but that line is never the bare gloss word. The other 97 entries settle the question on their own page, and in eleven of the sixteen glosses they settle it twice over, because the secondary senses carry by extension or figurative labels which name the core sense outright. The question is not open across the lexicon. It is unasked in 68 stubs and answered in the other 97.

Three languages, not two, keep a pair here: Gaillean, Middle Airananwe and Airananwe each hold two words for stone and two for bone. The first two were tagged by the rulings of 2026-08-21; the Airananwe pairs were not, and they are the only genuine question the section contains. Everywhere else one word covers the concept.

Dictionary:bone is the only one of the sixteen with an English lemma, and its translations box has already ruled: under gloss=one of the hard pieces that make up a skeleton it lists *jətew, ιετζου, eunduz, undes, *gjotonsra and all five Andusian words — the count sense — while leaving out the two material words ettuz and eces, and leaving out both Airananwe words entirely. For bone, then, which sense is settled; only whether to write it is not.

Twenty-one glosses can be corrected from the entries with no decision at all, and they are named gloss by gloss below. They follow the house form for a disambiguated gloss — one sense, near-synonyms joined by commas, a parenthetical only where the entry supplies one (calf (of cattle), stone (material), hair (of the head)) — and the ruling of 2026-08-04 that the comma joins and the semicolon explains, so whatever stands after a semicolon in a definition is description and has no place in a gloss.

The single question for the whole class: must every gloss in this class carry an explicit mass/count tag, or may a bare gloss stand where the language has only one word for the concept?

Option 1 — tag everything. All 165 glosses gain a tag. The box becomes self-describing, but for the 68 stubs the tag has to be invented: the gloss would assert a distinction the entry itself does not make.

Option 2 — default to both, tag only splits. A bare gloss means the word covers the substance and a piece of it alike; a tag is obligatory only where a language keeps two words — six lemmas, four already tagged, two pending the stone/bone ruling below. Nothing is invented, and the 21 corrections stand on their own. The cost is that a reader cannot tell "covers both" from "not yet examined".

Option 3 — fill the definitions and let the glosses follow. Treat the 68 one-line definitions as the defect, write them out as the Andusian entries already are, and derive each gloss from a real sense list. The largest job, but it removes the class of problem instead of notating around it.

Two notation points settle themselves under any option: the count tag is written "(a piece)" for stone and "(body part)" for bone, and a box read down a column cannot see that those are one category — one string should serve both; and the tag belongs in the gloss= field, which is what a translations box and the collision check both read.

hair

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*eɑkʲiəθ	hair	"hair, fur; a covering of hair"; "anything fine and thread-like" (by extension)
Ancient Hertic	ακειθι	hair	"hair" — nothing further
Gaillean	eaoklia	hair	"hair" — nothing further
Middle Airananwe	eyaankie	hair	"hair" — nothing further
Airananwe	yángly	hair	"hair" — nothing further
Proto-Andusian	*mamgob	hair	"hair (of the head)"; "locks"; "mane" (by extension)
Proto-North-Andusian	*mamgob	hair (of the head)	"hair (of the head)"; "locks"; "mane" (by extension)
Proto-South-Andusian	*mamgoʔ	hair (of the head)	"hair (of the head)"; "locks"; "mane" (by extension)
Proto-Unterian	*mamgobi	hair (of the head)	"hair (of the head)"; "locks"; "mane" (by extension)
Besoh	mamgo'	hair (of the head)	"hair (of the head)"; "locks"; "mane" (by extension)

Answerable from the entries. *mamgob is the only Andusian entry left bare while its own first definition reads "hair (of the head)" and all four of its reflexes carry that qualifier; its gloss should be hair (of the head). *eɑkʲiəθ should be hair, fur, which is the translating half of its first definition and rules out the single-strand reading by itself. That exposes a second split under one English word, and one that matters more here than mass against count: the Andusian five are head-hair, the Herta-Olaic word reaches fur, and the two sets are not interchangeable in one box. The four bare entries are the direct descendants of *eɑkʲiəθ (its Descendants tree runs ακειθι, eaoklia → eyaankie → yángly), so if the class ruling requires a tag they inherit "hair, fur"; no entry anywhere in the group lexicalises one hair, since "locks" and "mane" are collective and "anything fine and thread-like" is a marked extension.

stone

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*o:jig͡ɣ	stone	"stone, rock; a hard thing from the ground"; "a tool-stone"
Ancient Hertic	ωιγου	stone	"stone" — nothing further
Gaillean	oiat	stone (material)	"stone, as a material or substance"
Gaillean	oisie	stone (a piece)	"a stone; a piece of stone"
Middle Airananwe	oyte	stone (material)	"stone, as a material or substance"
Middle Airananwe	oysye	stone (a piece)	"a stone; a piece of stone"
Airananwe	oyt	stone	"stone" — nothing further
Airananwe	oyes	stone	"stone" — nothing further
Proto-Njeshan	*okju	stone	"stone" — nothing further
Proto-Andusian	*dʰɔɣ	stone	"stone, rock"; "hammerstone, hard tool" (by extension); "anything hard"; "steadfastness" and "a burden" (figurative)
Proto-North-Andusian	*dɔɣ	stone	"stone, rock"; "hammerstone, hard tool" (by extension)
Proto-South-Andusian	*hɔ	stone	"stone"; "quern, grindstone"
Proto-Unterian	*rɔɣi	stone	"stone"; "hammerstone, tool-stone"
Besoh	hoĥ	stone	"stone, rock"

Needs a ruling. Does the material-against-piece split ruled for Gaillean and Middle Airananwe continue into Airananwe oyt and oyes, the only other language in the section holding two words for the concept? The evidence stops exactly there: the Descendants tree on *o:jig͡ɣ branches oisie → oysye and oiat → oyte and goes no further, and the two Airananwe forms are among the only four words in this entire section with no etymology at all — every other Airananwe entry in the sixteen glosses has one. If the split survived, the glosses write themselves on shape alone (oyt "stone (material)" after oyte, oyes "stone (a piece)" after oysye) and the four missing etymologies can be filled in the same pass; if it did not, the Dictionary holds two undifferentiated synonyms needing either a fresh distinction or a merge, and Airananwe becomes the one place in the family where an inherited contrast was lost — a claim worth making deliberately rather than by silence. Nothing else in the table needs correcting: the Andusian entries mark their discrete tool senses by extension, so "stone" already names their core, and the enriched entries are consistent as they stand.

bone

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*jətew	bone	"bone; the hard frame"; "strength and structure" (by extension)
Ancient Hertic	ιετζου	bone	"bone" — nothing further
Gaillean	ettuz	bone (material)	"bone, as a substance or material"
Gaillean	eunduz	bone (body part)	"a bone; any of the hard pieces of a body, especially of a person"
Middle Airananwe	eces	bone (material)	"bone, as a substance or material"
Middle Airananwe	undes	bone (body part)	"a bone; any of the hard pieces of a body, especially of a person"
Airananwe	ytsis	bone	"bone" — nothing further
Airananwe	undeis	bone	"bone" — nothing further
Proto-Njeshan	*gjotonsra	bone	"bone" — nothing further
Proto-Andusian	*famrek	bone	"bone"; "the hard core of anything" (by extension); "frame"; "the kernel or gist" (figurative); "the irreducible part"
Proto-North-Andusian	*famrek	bone	"bone"; "frame, framework"
Proto-South-Andusian	*famreʔ	bone	"bone"; "frame"
Proto-Unterian	*famreki	bone	"bone"; "the father's line"
Besoh	famre'	bone	"bone"; "frame"

Needs a ruling. The same question as stone, in the same language: do Airananwe undeis and ytsis continue the split their Middle Airananwe ancestors carry, undes "bone (body part)" and eces "bone (material)"? It is one decision, not two, since it is one language and one inheritance — and *jətew's Descendants section already annotates the Gaillean pair in prose, "bone as a material" against "a bone of the body", so the distinction is recorded upstream and simply stops at Middle Airananwe. Everything else in this table is already answered off-page: Dictionary:bone places *jətew, ιετζου, *gjotonsra, *famrek, *famrek, *famreʔ, *famreki and famre' in the box for "one of the hard pieces that make up a skeleton", which is the count sense, and picks eunduz and undes rather than ettuz and eces. Those eight glosses are therefore correct as they stand, and become "bone (body part)" only if the class ruling calls for the tag; what they must not become is "bone; frame", since the frame senses are marked extensions and the box has already placed the words.

wood

Language	Word	Current gloss	What the entry's own definitions say
Gaillean	trolliquu	wood	"wood" — nothing further
Middle Airananwe	trolyiqe	wood	"wood" — nothing further
Airananwe	trolsque	wood	"wood" — nothing further
Proto-Andusian	*sim	wood	"wood, timber"; "firewood"; "material, stuff" (by extension)
Proto-North-Andusian	*sim	wood	"wood, timber"; "firewood"; "material, stuff" (by extension)
Proto-South-Andusian	*sim	wood	"wood"; "timber"
Proto-Unterian	*simi	wood	"wood, timber"; "firewood"; "material, stuff" (by extension)
Besoh	sim	wood	"wood"; "timber"

Answerable from the entries. No gloss needs changing, and the reason is worth recording: every attested sense is the substance, no entry reaches the woodland sense that makes bare wood dangerous in English, and the corpus keeps that sense in separate lemmas anyway — a forest word in all eight languages, and a firewood word (annawztrolle → annoztrol → annostrol) in the three whose entries are bare. If the class ruling requires a tag, "wood (material)" serves all eight. Do not fold "timber" into the gloss of *sim or sim, tempting though their second definition line makes it: both languages carry a separate lemma *simu / simu glossed "timber", and duplicating it would manufacture exactly the collision this page exists to clear. The near-homophony of sim "wood" with ʃim / śim "tree" in the same two languages is row 33's business, not this section's.

fruit

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*k͡xʲalɛt	fruit	"fruit; the edible growth of a plant"; "yield, produce" (by extension)
Ancient Hertic	αλπυτι	fruit	"fruit" — nothing further
Gaillean	ghaltu	fruit	"fruit" — nothing further
Middle Airananwe	ghaltye	fruit	"fruit" — nothing further
Airananwe	hatyu	fruit	"fruit" — nothing further
Proto-Andusian	*tsukze	fruit	"fruit"; "ripe produce"; "yield, gain, reward" (by extension); "consequence, outcome" (figurative)
Proto-North-Andusian	*tsuxze	fruit	"fruit"; "harvest"
Proto-South-Andusian	*tsukze	fruit	"fruit"; "ripe produce"; "yield, gain, reward" (by extension)
Proto-Unterian	*suze	fruit	"fruit"; "yield, reward"
Besoh	cuzze	fruit	"fruit"; "ripe produce"; "yield, gain, reward" (by extension)

Answerable from the entries. No gloss needs changing. Every sense beyond the headword is collective or abstract — produce, harvest, yield, outcome — and the abstract ones are labelled extensions, so "fruit" already names the core; the only entry that explains itself, *k͡xʲalɛt, calls it "the edible growth of a plant", which is the substance reading. Nothing in the group means one fruit as a single object, so if the class ruling calls for a tag the produce sense is the one the entries support, and it should not be invented in the other direction.

grass

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*dβəbutʃ	grass	"grass, low green growth, ground cover"; "fodder"
Ancient Hertic	φουβουξ	grass	"grass" — nothing further
Gaillean	dvuca	grass	"grass" — nothing further
Middle Airananwe	dhuce	grass	"grass" — nothing further
Airananwe	dhuse	grass	"grass" — nothing further
Proto-Andusian	*bʰat	grass	"grass"; "herbs, low green growth"; "pasture, thatch" (by extension)
Proto-North-Andusian	*bɑt	grass	"grass"; "pasture"
Proto-South-Andusian	*haʔ	grass	"grass"; "herbs, low green growth"; "pasture, thatch" (by extension)
Proto-Unterian	*wati	grass	"grass"; "thatch"
Besoh	ha'	grass	"grass"; "herbs, low green growth"; "pasture, thatch" (by extension)

Answerable from the entries. No gloss needs changing. Pasture, thatch, fodder and ground cover are all the undifferentiated stuff, and nothing here means a blade of grass or a species of grass, which are the two count readings English allows — so the mass reading is settled and the tag, if the class ruling wants one, is not a guess. The point the table does raise is extent rather than number: the Andusian words reach herbs and low growth generally, so "grass" is narrow for them rather than merely underspecified, and *bɑt and *wati leave their second sense unlabelled where their three relatives mark it by extension.

rope

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*ᵐbxo:klə	rope	"rope, cord; a twisted line"; "binding and constraint" (by extension)
Ancient Hertic	μιωγιλ	rope	"rope" — nothing further
Gaillean	mbokiluz	rope	"rope" — nothing further
Middle Airananwe	mbokilis	rope	"rope" — nothing further
Airananwe	bockils	rope	"rope" — nothing further
Proto-Andusian	*ɔᵑg	rope	"rope, cord"; "vine, liana"; "a bond, an obligation" (figurative)
Proto-North-Andusian	*oᵑg	rope	"rope"; "tether"
Proto-South-Andusian	*ɔŋ	rope	"rope, cord"; "vine, liana"
Proto-Unterian	*ɔŋu	rope	"rope, cord"; "vine, liana"
Besoh	one	rope	"rope, cord; vine, liana" (one line)

Answerable from the entries. No gloss needs changing, but this is the one gloss in the class whose evidence points the other way, and that is the finding. A twisted line, a tether, a vine, a bond: every added sense is a discrete object, so these are the object word, not a name for rope as a material — which is an argument against any blanket "the class is the mass sense" default. One housekeeping matter, not a gloss question: one puts "vine, liana" after a semicolon on the same line where its four relatives give it a numbered sense of its own, and under the 2026-08-04 ruling a semicolon introduces description, not a second translation.

sand

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*ʃuʊnɛwɹ	sand	"sand, grit; loose dry ground" (one line)
Ancient Hertic	ςεριου	sand	"sand" — nothing further
Gaillean	ceunetra	sand	"sand" — nothing further
Middle Airananwe	cunetaa	sand	"sand" — nothing further
Airananwe	suntá	sand	"sand" — nothing further
Proto-Andusian	*lak	sand	"sand"; "grit"; "the beach" (by extension)
Proto-North-Andusian	*lak	sand	"sand"; "grit"; "the beach" (by extension)
Proto-South-Andusian	*laʔ	sand	"sand"; "grit"; "the beach" (by extension)
Proto-Unterian	*laki	sand	"sand"; "grit"; "the beach" (by extension)
Besoh	la'	sand	"sand"; "grit"; "the beach" (by extension)

Answerable from the entries. No gloss needs changing, and on the mass/count axis this gloss barely belongs in the class: a grain of sand appears nowhere, and English sand is not a count word in the relevant sense either. The five Andusian entries do carry a place sense, "the beach", which the gloss suppresses — but it is labelled by extension in every one of them, so under the core-meaning rule it belongs in the definitions and stays out of the gloss.

ash

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*əmo:ɑ	ash	"ash; the leavings of fire, fine grey dust" (one line)
Ancient Hertic	εμβοι	ash	"ash" — nothing further
Gaillean	emmora	ash	"ash" — nothing further
Middle Airananwe	emere	ash	"ash" — nothing further
Airananwe	emmre	ash	"ash" — nothing further
Proto-Njeshan	*bulʃpə	ash	"ash" — nothing further
Proto-Andusian	*ɲoquᵐb	ash	"ash, cinders"; "the hearth" (by extension); "mourning-dust"
Proto-North-Andusian	*ɲoʔuᵐb	ash	"ash, cinders"; "the hearth" (by extension)
Proto-South-Andusian	*ɲoɢum	ash	"ash, cinders"; "the hearth" (by extension)
Proto-Unterian	*noʔumu	ash	"ash"; "the hearth's ash"; "mourning-dust"
Besoh	ñoǵum	ash	"ash, cinders"; "the hearth" (by extension)

Answerable from the entries. Cinders is a plural of the same stuff, not a count word, and no entry offers one, so mass against count does not arise. The gloss has a different defect: bare ash is a homograph in English and a box gives no clue that the tree is not meant. Four entries supply the fix in their own first line, and their glosses should become ash, cinders — *ɲoquᵐb, *ɲoʔuᵐb, *ɲoɢum and ñoǵum. The remaining seven, including *noʔumu whose own first line is bare, stay as they are unless the class ruling extends "ash, cinders" to them.

dust

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*xɛʃtɛʃ	dust	"dust; fine dry matter, ash-like powder" (one line)
Ancient Hertic	εςθισι	dust	"dust" — nothing further
Gaillean	khest	dust	"dust" — nothing further
Middle Airananwe	kyest	dust	"dust" — nothing further
Airananwe	khest	dust	"dust" — nothing further
Proto-Njeshan	*fuŋgeŋ	dust	"dust" — nothing further
Proto-Andusian	*buɢ	dust	"dust"; "powder"; "dry loose earth"
Proto-North-Andusian	*buɣ	dust	"dust"; "powder"; "dry loose earth"
Proto-South-Andusian	*buʔ	dust	"dust"; "powder"; "dry loose earth"
Proto-Unterian	*bugi	dust	"dust"; "powder"; "dry loose earth"
Besoh	bu'	dust	"dust"; "powder"; "dry loose earth"

Answerable from the entries. Mass throughout, no count sense anywhere, no gloss needs changing. One correction to the record: the senses "near, close", "beside" and "imminent" that appear against *buʔ and bu' in a flat export of these pages are not senses of the dust noun — both pages already split Etymology 1 (noun, from *buɢ "dust") from Etymology 2 (adjective, from *but "near"), and the Lexemes table holds them as two rows. The homograph is handled; there is nothing to raise.

salt

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*tɛwɛɹɲɔ	salt	"salt (the substance that preserves)"; "the salt taste"
Ancient Hertic	θερνουξ	salt	"salt" — nothing further
Gaillean	etrenuz	salt	"salt" — nothing further
Middle Airananwe	etrengis	salt	"salt" — nothing further
Airananwe	etrenge	salt	"salt" — nothing further
Proto-Njeshan	*inur	salt	"salt" — nothing further
Proto-Andusian	*izvo	salt	"salt"; "savor, taste" (by extension); "the sea's taste"; "the sea itself" (by metonymy); "wit, savor of speech" (figurative)
Proto-North-Andusian	*izvo	salt	"salt"; "savor, taste" (by extension)
Proto-South-Andusian	*izvo	salt	"salt"; "brine for preserving"
Proto-Unterian	*ivo	salt	"salt"; "savor, taste" (by extension)
Besoh	izvo	salt	"salt"; "brine for preserving"

Answerable from the entries. *tɛwɛɹɲɔ states the answer in its own first line and should carry it: gloss salt (the substance), the entry's parenthetical "(the substance that preserves)" cut to the length a gloss wants. Nothing in the group is a count word, so no other gloss changes: the taste, the brine and the sea are extensions in every entry that has them and stay in the definitions. Two small things fall out of the table: *izvo and izvo leave "brine for preserving" unlabelled where their relatives label the comparable sense, and four definition lines across three entries spell savor in the American form where the wiki's prose is British-spelt.

fat

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*p͡ɸa:z	fat	"fat, grease, richness"; "abundance" (by extension)
Ancient Hertic	βαζου	fat	"fat" — nothing further
Gaillean	bhaeuz	fat	"fat" — nothing further
Middle Airananwe	bhais	fat	"fat" — nothing further
Airananwe	vhais	fat	"fat" — nothing further
Proto-Njeshan	*rjolo	fat	"fat" — nothing further
Proto-Andusian	*cegb	fat	"fat, grease"; "oil"; "richness, abundance" (by extension); "the best part of anything" (figurative)
Proto-North-Andusian	*cegb	fat	"fat, grease"; "butterfat"
Proto-South-Andusian	*ʃeʔ	fat	"fat, grease"; "pressed oil"
Proto-Unterian	*sebi	fat	"fat"; "the best part of anything"
Besoh	śe'	fat	"fat, grease"; "pressed oil"

Answerable from the entries. The mass/count axis is inert — no entry means "a fat" — but bare fat carries a worse ambiguity that the definitions already dissolve: in English it is an adjective before it is a noun, and a box showing "fat" against a column of words says nothing about which. Five entries already say "grease", which cannot be read as an adjective, and their glosses should become fat, grease: *p͡ɸa:z, *cegb, *cegb, *ʃeʔ and śe'. *sebi does not say it and is left alone. The five entries actually at risk are the bare ones — bhaeuz, bhais, vhais, βαζου and Proto-Njeshan *rjolo, four Herta-Olaic and one Njeshan — and "fat, grease" would settle them at no interpretive cost if the class ruling allows it.

meat

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*oʊjɔʒ	meat	"flesh; the edible substance of a body"; "body, substance" (by extension)
Ancient Hertic	ουζαν	meat	"meat" — nothing further
Gaillean	oeuz	meat	"meat" — nothing further
Middle Airananwe	oesis	meat	"meat" — nothing further
Airananwe	oeshy	meat	"meat" — nothing further
Proto-Njeshan	*pispa	meat	"meat" — nothing further
Proto-Andusian	*sits	meat	"meat, flesh"; "the body" (by extension); "the soft flesh of fruit"
Proto-North-Andusian	*sits	meat	"meat, flesh"; "the body" (by extension)
Proto-South-Andusian	*sits	meat	"meat, flesh"; "the body" (by extension)
Proto-Unterian	*sisi	meat	"meat, flesh"; "the body" (by extension)
Besoh	sic	meat	"meat, flesh"; "the body" (by extension)

Answerable from the entries. *oʊjɔʒ is glossed "meat" while its own first definition begins "flesh", and under the 2026-08-04 ruling the translating half of that line is the single word before the semicolon; its gloss should be flesh. The change is not cosmetic. Its four reflexes are all glossed "meat", and the same three daughters keep a separate flesh word (thembat → thembte → thebte), so the proto is the wide word and the daughters narrowed — and under the core-meaning rule the corrected gloss also keeps *oʊjɔʒ out of an English meat box, where it does not belong. If the operator instead means the proto as meat, it is the definition line that has to change, not the gloss. The Andusian five need nothing: "meat, flesh" is their first line and "the body" is marked an extension.

skin

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*əkgɛɪrdβ	skin	"skin, hide; the outer covering of a body" (one line)
Ancient Hertic	κεγηρτι	skin	"skin" — nothing further
Gaillean	ennerd	skin	"skin" — nothing further
Middle Airananwe	ender	skin	"skin" — nothing further
Airananwe	enner	skin	"skin" — nothing further
Proto-Andusian	*abʰfiq	skin	"skin, hide"; "peel, rind, shell"; "surface" (by extension); "outward seeming, appearance" (figurative)
Proto-North-Andusian	*abfiʔ	skin	"skin, hide"; "peel, rind"; "surface" (by extension)
Proto-South-Andusian	*aːfiʔ	skin	"skin, hide"; "peel, rind"; "surface" (by extension)
Proto-Unterian	*awfiʔi	skin	"skin, hide"; "peel, rind"; "surface" (by extension)
Besoh	äfi'	skin	"skin, hide"; "peel, rind"; "surface" (by extension)

Answerable from the entries. Six entries open with "skin, hide" and their glosses should say so — skin, hide for *əkgɛɪrdβ, *abʰfiq, *abfiʔ, *aːfiʔ, *awfiʔi and äfi'. Hide is the separable piece and skin the covering, so these words cover both and the mass/count question answers itself without a tag. The sense a bare gloss really loses is "peel, rind": that is a plant-against-animal distinction, not a mass-against- count one, it is a numbered sense rather than a labelled extension, and it stays in the definitions.

snow

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*ɔwb͡βiətβ	snow	"snow, snowfall"; "the white cover of winter"
Ancient Hertic	ωφητου	snow	"snow" — nothing further
Gaillean	orbhita	snow	"snow" — nothing further
Middle Airananwe	erbhiite	snow	"snow" — nothing further
Airananwe	yrbit	snow	"snow" — nothing further
Proto-Andusian	*muᵐb	snow	"snow, frost; the white of the high peaks"; "hoar, white crust; any white thing" (by extension)
Proto-North-Andusian	*muᵐb	snow	"snow"; "the white of the peaks"

Answerable from the entries. Mass throughout — "snowfall" is an event, not a piece — and one gloss under-reports its own entry: *muᵐb should read snow, frost, the translating half of its first line, since frost is a second translation and not a description of snow. The seven rows are not the whole story and the gap in them is not a gap: the etymon has reflexes in the four languages missing from the table, glossed elsewhere — *mum "hail" and mium "hail" (the pair ruled synonyms of the rɔts words on 2026-08-21) and *mumu "sea-foam". Only Proto-Njeshan has no word in this field at all. The southern branch turned a snow word into a hail word, which is a fact about the family, not thin evidence.

ice

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*kɪɑnɪntʃ	ice	"ice, frozen water; hardness from cold" (one line)
Ancient Hertic	χαινταν	ice	"ice" — nothing further
Gaillean	qanitzh	ice	"ice" — nothing further
Middle Airananwe	qaancidye	ice	"ice" — nothing further
Airananwe	qánide	ice	"ice" — nothing further
Proto-Andusian	*rɔts	ice	"hail, hailstone"; "ice, known chiefly from the high peaks"; "anything glassy and hard" (by extension)
Proto-North-Andusian	*rots	ice	"ice"; "hail"

Answerable from the entries. *rɔts is the clearest miss in the section: it is glossed "ice", but its own first sense is "hail, hailstone" and its second marks ice as known chiefly from the high peaks. Its gloss should be hail — and the entry is not the only witness, because every daughter's etymology line already cites it as "hail": *rots, *rɔts, *rɔsi and roc, the last three of which are themselves glossed "hail" and so fall outside this table. The table is short for that reason, not because the languages lack the word; only Proto-Njeshan is genuinely absent. *rots keeps ice first and needs nothing. The one count word anywhere in the section is "hailstone", and it belongs to that hail word rather than to ice, which is mass in every entry that has it.

Where the word covers a different extent

hand

Language	Word	Current gloss	What the entry's own definition says
Proto-Andusian	*ᵐbu	hand	"the hand and forearm"; then grasp, possession; then power, control. Explicitly past the wrist.
Proto-Herta-Olaic	*βiɑɹɛʃ	hand	"hand; the grasping part"; then holding, possession, a handful. The hand proper.
Gaillean	variza	hand	"hand", nothing further — but kubhe "arm" is a separate lemma.
Middle Airananwe	varize	hand	"hand", nothing further; kuuph "arm" beside it.
Airananwe	varris	hand	"hand", nothing further; couphe "arm" beside it.
Ancient Hertic	βαρζισι	hand	"hand", nothing further; no arm lemma in the language.
Proto-Njeshan	*mtjaxŋ	hand	"hand", nothing further; no arm lemma in the language.

Answerable from the entries. The two words that describe themselves describe different body parts, so the answer is that they are not the same: *ᵐbu runs up the forearm and *βiɑɹɛʃ stops at the grasping part. Set *ᵐbu to hand and forearm and *βiɑɹɛʃ to hand, the grasping part. The first of those is housekeeping rather than a decision: every reflex of *ᵐbu already carries "hand and forearm" in its gloss field — *ᵐbu, *bu, *mu and Besoh bu — and all four of those entries cite the etymon in their etymology as *ᵐbu ("hand and forearm"), so the proto is the only entry in its own family out of step with itself. The five bare words keep hand: Gaillean, Middle Airananwe and Airananwe each hold a separate arm lemma, which bounds their reflexes from outside, and although βαρζισι and *mtjaxŋ have nothing beside them, neither entry claims more than the hand. Two consequences follow for the English entry rather than for the glosses. *ᵐbu is listed in the translations box of Dictionary:hand, which is headed "the part of the arm below the wrist" and therefore cannot hold a word that includes the forearm; and varris is absent from that box although the entry exists.

earth

Language	Word	Current gloss	What the entry's own definition says
Proto-Herta-Olaic	*ihe:x	earth	One sense line: "ground, soil, earth; the land underfoot — not the planet". Ground and soil together, the planet ruled out in as many words.
Gaillean	iat	earth	"earth", nothing further — but eta "soil", sicne "land" and contazu "world" are separate lemmas, none of them related to it.
Middle Airananwe	ieeth	earth	"earth", nothing further; ete "soil", siicen "land", tyonchus "world" beside it.
Airananwe	éth	earth	"earth", nothing further; ette "soil", sichen "land", tonchous "world" beside it.
Ancient Hertic	ιρηγαν	earth	"earth", nothing further; no soil, ground, land or world lemma in the language.
Proto-Andusian	*tse	earth	"earth, soil"; "the ground as substance"; then, by extension, the world, the land. Marked a coordinate term of *lem "ground".
Proto-North-Andusian	*tse	earth	As Proto-Andusian: earth and soil, the ground as substance, and by extension the world.
Proto-South-Andusian	*tse	earth	As Proto-Andusian.
Proto-Unterian	*se	earth	As Proto-Andusian.
Besoh	ce	earth	As Proto-Andusian.
Proto-Njeshan	*muto	earth	"earth", nothing further; no soil, ground, land or world lemma in the language.

Answerable from the entries. Two extents are on record and they differ over the substance, not over the surface. The Andusian word is the substance: "earth, soil", "the ground as substance", extended to the world, with *lem "ground" marked beside it as its coordinate term, so the surface is *lem's job and not its own. The Herta-Olaic etymon covers ground and soil in one line and strikes out the planet. Set *tse, *tse, *tse, *se and ce to soil; set *ihe:x to ground, soil. iat, ieeth and éth take ground — not on their own wording, which is bare, but because their languages have already given the rest away: each holds an unrelated lemma for soil, for land and for world, so of the etymon's range only "the land underfoot" is left for them, and no lemma anywhere in the corpus is glossed "floor", "surface" or "planet" for that reading to collide with. ιρηγαν and *muto keep earth unchanged: nothing in Ancient Hertic or Proto-Njeshan competes with either, and neither entry says more, so there is nothing on record to correct them to. The consequence falls on Dictionary:earth, whose sole sense line reads "Soil; the loose ground underfoot, in which plants grow": under the above it is listing the wrong Herta-Olaic words, since eta, ete and ette belong in that box and iat and ieeth belong under a surface sense. éth is absent from the box either way.

ground

Language	Word	Current gloss	What the entry's own definition says
Proto-Andusian	*lem	ground	"ground, floor"; "the low place"; "bottom"; then foundation, basis. Marked a coordinate term of *tse "earth".
Proto-North-Andusian	*lem	ground	"ground"; "the tilled field". Narrowed to worked land; floor, low place and bottom are all gone.
Proto-South-Andusian	*lem	ground	"ground, floor"; "the low place"; "bottom".
Proto-Unterian	*lemi	ground	"ground, floor"; "the bottom, also of the sea". Extended to the seabed.
Besoh	lem	ground	"ground, floor"; "the low place"; "bottom". Inherited through *lem.

Answerable from the entries. All five state their own extent, and two of them are not what the other three are. Set *lem to ground, tilled field and *lemi to ground, floor, sea bottom; *lem, *lem and lem take ground, floor. Taken with the earth changes above this settles the earth-against-ground question outright, and settles it from the corpus rather than by ruling: within Andusian the two words are marked coordinate terms of each other, *tse being the substance and *lem the surface and the low place. No Dictionary:ground entry exists to anchor the English side; when one is written, "the surface underfoot" is the sense these five and the three Herta-Olaic words moved to ground above all answer to.

tree

Language	Word	Current gloss	What the entry's own definition says
Proto-Herta-Olaic	*dwɑl	tree	"tree, standing wood; the living plant and its timber undifferentiated". The only word in the group that says it covers both.
Gaillean	dwalia	tree	"tree", nothing further — but trolliquu "wood" is a separate lemma.
Middle Airananwe	diaalie	tree	"tree", nothing further; trolyiqe "wood" beside it.
Airananwe	tiále	tree	"tree", nothing further; trolsque "wood" beside it.
Ancient Hertic	αλον	tree	"tree", nothing further; no wood lemma in the language, though σαττουξ "forest" exists.
Proto-Andusian	*ciᵐb	tree	"tree"; then trunk, timber; then a line of descent. Marked a coordinate term of *sim "wood".
Proto-North-Andusian	*ciᵐb	tree	"tree"; then trunk, timber. *sim "wood" beside it.
Proto-South-Andusian	*ʃim	tree	"tree"; "timber-tree". *sim "wood" and *simu "timber" beside it.
Proto-Unterian	*simu	tree	"tree"; then trunk, timber. *simi "wood" beside it.
Besoh	śim	tree	"tree"; "timber-tree". sim "wood", simu "timber" and śimu "forest" beside it.

Answerable from the entries. One word covers the plant and its material together and says so: set *dwɑl to tree, standing wood. Every other word in the group keeps tree, because every other language holds a separate lemma for wood — *sim and its reflexes throughout Andusian, and trolliquu, trolyiqe, trolsque in the three Herta-Olaic daughters, which shows the merge in *dwɑl did not survive into them. αλον keeps "tree" on the same footing, with the caveat that Ancient Hertic has no wood lemma to bound it and its entry claims no more than the tree. The near-homophony needs no ruling either, because the etymologies already account for it: *ʃim, *simu and śim all give *ciᵐb "tree" as their etymon while *sim, *simi and sim give *sim "wood", so the forms converged by regular change and the lemmas stayed apart. Besoh is where the convergence piles up hardest and is still consistent — śim "tree", śimu "forest" (from *ciᵐbuɣ, and also the plural of śim), sim "wood" and simu "timber" (from *sim plus a suffix, and also the plural of sim) — four lemmas kept distinct only by the s/ś contrast. tiále is missing from the Dictionary:tree box. The wood half of this row is a mass-against-count question and belongs with row 29, not here.

breast

Language	Word	Current gloss	What the entry's own definition says
Proto-Herta-Olaic	*pvfo:ʃə	breast	"chest; the front of the torso"; then "the breast that feeds". The gloss names the second sense, not the first.
Gaillean	phocta	breast	"breast", nothing further — but louwat "chest" is a separate lemma, with its etymology undetermined and so not a doublet.
Middle Airananwe	phoce	breast	"breast", nothing further; luuyaazy "chest" beside it.
Airananwe	phosse	breast	"breast", nothing further; luáj "chest" beside it.
Ancient Hertic	πηφουζι	breast	"breast", nothing further; no chest lemma in the language.
Proto-Njeshan	*krapɛstə	breast	"breast", nothing further; no chest lemma in the language.
Proto-Andusian	*dʰirpeᵐb	breast	"breast"; "chest"; then nursing, mother's milk. Both readings, as separate senses.
Proto-North-Andusian	*derpeᵐb	breast	As Proto-Andusian: breast, chest, nursing.
Proto-South-Andusian	*hirpem	breast	As Proto-Andusian.
Proto-Unterian	*rirpemu	breast	As Proto-Andusian.
Besoh	hirpem	breast	As Proto-Andusian.

Answerable from the entries. The row's "unqualified" is wrong: six of the eleven entries answer the question on their own pages, and the rest are bounded from outside. *pvfo:ʃə leads with the chest and puts the feeding breast second, so its gloss records the wrong sense — set it to chest, breast. The five Andusian words list breast and chest as separate senses and take breast, chest. phocta, phoce and phosse take breast (the mammary gland): each of those three languages holds louwat, luuyaazy or luáj "chest" as a lemma of its own, unrelated to the etymon, so the chest is occupied and what these continue is the etymon's second sense. πηφουζι and *krapɛstə keep breast unchanged, having no second lemma to bound them and no wording of their own to go on. No Dictionary:breast entry exists; when one is written it needs two sense lines, and only *pvfo:ʃə and the five Andusian words may stand in both boxes.

head

Language	Word	Current gloss	What the entry's own definition says
Proto-Herta-Olaic	*əɲib͡βeə	head	"head; the top of a person"; then "the front, the chief, the beginning".
Gaillean	enibhia	head	"head", nothing further.
Middle Airananwe	enibhie	head	"head", nothing further.
Airananwe	nibhe	head	"head", nothing further.
Ancient Hertic	μενηβι	head	"head", nothing further.
Proto-Njeshan	*kokint	head	"head", nothing further.
Proto-Andusian	*ɣa	head	"head"; "top, summit"; then "chief, foremost"; then the beginning of anything. All three readings.
Proto-North-Andusian	*ɣa	head	"head"; "chief, headman". Kept the leader and dropped the top.
Proto-South-Andusian	*ɣa	head	"head"; "the source of a river". Neither leader nor top.
Proto-Unterian	*ɣa	head	"head"; "top". Kept the top and dropped the leader.
Besoh	ĝa	head	"head"; "the source of a river". As Proto-South-Andusian, from which it is inherited.

Answerable from the entries. Four reflexes of one Proto-Andusian etymon each kept a different extension and the identical gloss now hides it, which is the sharpest case in this section: *ɣa went to the leader, *ɣa and ĝa to the river source, *ɣa to the top and no further. Set *ɣa to head, top, chief, *ɣa to head, chief, *ɣa and ĝa to head, river source, *ɣa to head, top, and *əɲib͡βeə to head, chief. The North Andusian reading is corroborated twice over: *ɣaɣe "chief, lord" and *ɣauɣ "council" are both formations on *ɣa that survive in that branch alone. The five that say only "head" keep head and mean the body part, since nothing on their pages claims more; note that this is what their entries say rather than an argument from the rest of the corpus, which holds no lemma glossed "leader" or "top" either way — the only such lemma is *ɣaɣe, a derivative of *ɣa itself. The consequence falls on Dictionary:head, which does not yet exist: it needs three sense lines, and only the six words above may enter the chief, top and river-source boxes.

ear

Language	Word	Current gloss	What the entry's own definition says
Proto-Herta-Olaic	*oʊjwə	ear	"ear, hearing" as one sense; then "attention and obedience". Organ and faculty undifferentiated.
Gaillean	oelwuz	ear	"ear", nothing further.
Middle Airananwe	yelue	ear	"ear", nothing further.
Airananwe	yelue	ear	"ear", nothing further. Spelled the same as its Middle Airananwe parent.
Ancient Hertic	ειουξ	ear	"ear", nothing further.
Proto-Njeshan	*nɛrʃot	ear	"ear", nothing further.
Proto-Andusian	*ɣɔix	ear	"ear"; then, by extension, "hearing, attention, heed". The faculty is a marked second sense, not part of the first.
Proto-North-Andusian	*ɣoix	ear	As Proto-Andusian.
Proto-South-Andusian	*ɣɔi	ear	As Proto-Andusian.
Proto-Unterian	*ɣɔixi	ear	As Proto-Andusian.
Besoh	ĝoi	ear	As Proto-Andusian.

Answerable from the entries. The ear-of-grain reading is not in the corpus at all: every one of the eleven definitions names the organ or the faculty of hearing, and no lexeme in the table is glossed for grain or any cereal, so the question this row asks does not arise and ear stands for ten of the eleven. The difference that is genuinely on record lies elsewhere — *oʊjwə folds organ and faculty into a single first sense where the five Andusian entries keep hearing as a marked extension — so set *oʊjwə to ear, hearing and leave the rest. Nothing in the corpus is glossed "hearing" as a noun; Dictionary:hear carries that ground as a verb, with *g͡ɣije:ht and γιστω among its translations. When Dictionary:ear is written, "the organ of hearing" is the sense these eleven answer to.

Adjectives running on two axes

big

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*ɔjti:	big	one line: "big, large, much, important — size, quantity and significance undifferentiated"
Ancient Hertic	ωιθε	big	"big" — sole line
Gaillean	oitha	big	"big" — sole line
Middle Airananwe	otya	big	"big" — sole line
Airananwe	othay	big	"big" — sole line
Proto-Njeshan	*ŋbarisk	big	"big" — sole line
Proto-Andusian	*axgɛɣ	big	big, great / grown, full-sized / (by extension) important / (figurative) great in rank, senior
Proto-North-Andusian	*axgeɣ	big	big, great / great in rank
Proto-South-Andusian	*axgɛ	big	big, great / grown, full-sized / (by extension) important
Proto-Unterian	*agɛɣi	big	big / full-grown
Besoh	agge	big	big, great / grown, full-sized / (by extension) important

Answerable from the entries. Six of the eleven state their own range. Set *ɔjti: to big (in size, quantity or importance); *axgɛɣ to big, great (in size or growth); important, senior in rank; *axgeɣ to big, great (in size); great in rank; *axgɛ and agge to big, great (in size or growth); important; *agɛɣi to big (in size); full-grown. The daughters have already divided the proto's four senses between them — *axgeɣ keeps rank but drops growth and importance, *agɛɣi keeps growth and drops both — so these are five distinct glosses, not one repeated.

Needs a ruling. Do the five bare entries ωιθε, oitha, otya, othay and *ŋbarisk cover importance as well as physical size? If the Herta-Olaic four are read as continuing *ɔjti:, they take big (in size or importance), the branch matches the Andusian pattern, and no separate root for "important" is owed; if they are read narrowly they take big (in size), importance becomes a lexical gap in the branch, and each entry should record that it has narrowed against its proto. *ŋbarisk is the harder half: Proto-Njeshan is itself a proto, so there is no parent entry to read down into it and the sense must be stated outright. The same shape of question recurs under all ten glosses below, and a single blanket policy on whether a bare daughter inherits its proto's full range would settle the Herta-Olaic rows everywhere at once — but it would not touch the four Proto-Njeshan rows (here and under old, new and dull), and the substance differs gloss by gloss, so ruling one by one remains open.

long

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*oʊɹmɪh	long	one line: "long; extended in space or time"
Ancient Hertic	ρυμε	long	"long" — sole line
Gaillean	oembia	long	"long" — sole line
Middle Airananwe	oouebie	long	"long" — sole line
Airananwe	oubiey	long	"long" — sole line
Proto-Andusian	*ᵑgaha	long	long / tall / stretched out / (by extension) lasting, enduring (of time) / (figurative) wearisome, drawn-out
Proto-North-Andusian	*ᵑgaa	long	long / tall / stretched out / (by extension) lasting, enduring (of time)
Proto-South-Andusian	*gaha	long	long / tall / stretched out / (by extension) lasting, enduring (of time)
Proto-Unterian	*ŋaha	long	long / tall / stretched out / (by extension) lasting, enduring (of time)
Besoh	gä	long	long / tall / stretched out / (by extension) lasting, enduring (of time)

Answerable from the entries. Set *oʊɹmɪh to long (in space or time); *ᵑgaha to long (in space or time); tall; drawn-out; and *ᵑgaa, *gaha, *ŋaha and gä to long (in space or time); tall. "Stretched out" belongs with sense 1 and needs no separate slot in the gloss. The vertical sense is attested only in Andusian and should not be carried elsewhere.

Needs a ruling. Do the four bare Herta-Olaic words carry duration as well as spatial extent, as *oʊɹmɪh does? Inheriting gives ρυμε, oembia, oouebie and oubiey the gloss long (in space or time); reading them narrowly gives long (in space) and leaves duration to a root not yet coined. There is no Njeshan entry for this gloss, so the ruling covers four words only. Whichever way it goes should govern short too, or the antonym pair will be asymmetrical.

short

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*ɸəmnɛt	short	one line: "short, brief, not far — space and time together"
Ancient Hertic	ψεντε	short	"short" — sole line
Gaillean	femetta	short	"short" — sole line
Middle Airananwe	femethe	short	"short" — sole line
Airananwe	mettey	short	"short" — sole line
Proto-Andusian	*luse	short	short / low / brief (of time)
Proto-North-Andusian	*luse	short	short / low / brief (of time)
Proto-South-Andusian	*luse	short	short / low / brief (of time)
Proto-Unterian	*luse	short	short / low / brief (of time)
Besoh	luze	short	short / low / brief (of time)

Answerable from the entries. Set *ɸəmnɛt to short (in space or time), and the five Andusian entries *luse, *luse, *luse, *luse and luze to short (in space or time); low. The four proto-forms are not merely alike in spelling: they occupy four language sections of the one page Dictionary:luse, so the gloss is the only field that will tell their rows apart in a translations box. Note also that this set is the exact mirror of long — low answers tall, brief answers lasting — and that no reflex has narrowed.

Needs a ruling. Do ψεντε, femetta, femethe and mettey cover brevity of time as well as shortness in space? Inheriting gives short (in space or time); reading them narrowly gives short (in space). There is no Njeshan entry here either. This must match the ruling on long.

old

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*hɛkʃa	old	one line: "old, long-standing, aged — of things and of people alike"
Ancient Hertic	εξε	old	"old" — sole line
Gaillean	exa	old	"old" — sole line
Middle Airananwe	exa	old	"old" — sole line
Airananwe	exay	old	"old" — sole line
Proto-Njeshan	*brote	old	"old" — sole line
Proto-Andusian	*dʰukif	old	old / worn, used / elder, senior (of people) / (figurative) venerable / long-established
Proto-North-Andusian	*dokif	old	old / senior, of rank
Proto-South-Andusian	*hugif	old	old / worn, used / elder, senior (of people) / long-established
Proto-Unterian	*rukifi	old	old / venerable, elder
Besoh	hugif	old	old / worn, used / elder, senior (of people) / long-established

Answerable from the entries. Set *hɛkʃa to old (of things and of people); *dʰukif to old, worn, long-established; elder, venerable (of people); *dokif to old; senior, of rank; *hugif and hugif to old, worn, long-established; elder, senior (of people); *rukifi to old; venerable, elder. *dokif is the one Andusian reflex with neither the "worn, used" line nor a personal-age line — rank only — and its gloss should not be levelled with the others. *hugif and hugif likewise share a spelling and the single page Dictionary:hugif, with identical definitions, so only the language field distinguishes them.

Needs a ruling. Do the bare εξε, exa, exa, exay and *brote apply to people as well as to things? Inheriting *hɛkʃa's explicit "of things and of people alike" gives the Herta-Olaic four old (of things and of people); restricting them gives old (of things) and obliges a separate word for "elderly" in each. *brote again falls outside the inheritance question, having no parent entry, and needs the sense stated outright.

new

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*βo:xʒɑ	new	new, fresh, recent / (by extension) young and unfamiliar
Ancient Hertic	φοσε	new	"new" — sole line
Gaillean	vozwa	new	"new" — sole line
Middle Airananwe	fozya	new	"new" — sole line
Airananwe	fosay	new	"new" — sole line
Proto-Njeshan	*woge	new	"new" — sole line
Proto-Andusian	*ᵐbakp	new	new, fresh / young / raw, unused / (figurative) untried, inexperienced
Proto-North-Andusian	*ᵐbakp	new	new, fresh / young / raw, unused
Proto-South-Andusian	*baʔ	new	Etymology 2 only: new, fresh / young / raw, unused
Proto-Unterian	*mapi	new	new, fresh / young / raw, unused
Besoh	beza	new	new, fresh — sole line

Answerable from the entries. Set *βo:xʒɑ to new, fresh, recent; young, unfamiliar; *ᵐbakp to new, fresh; young; raw, unused; untried; *ᵐbakp, *baʔ and *mapi to new, fresh; young; raw, unused; beza to new, fresh. The Proto-Andusian and Proto-North-Andusian forms are spelt alike and share the page Dictionary:ᵐbakp; the parent's fourth, figurative sense is the whole of the difference, so the gloss must show it rather than substitute it. One warning about *baʔ: that page carries two etymologies, the first an unrelated adjective from Proto-Andusian bagb meaning "bad, harmful, ugly, ill-omened", the second the "new" lexeme from *ᵐbakp. Any sweep that reads definition lines off the page rather than off the sense block will merge the two and appear to show a word meaning both "bad" and "new".

Needs a ruling. Do the bare φοσε, vozwa, fozya, fosay and *woge carry novelty and unfamiliarity as well as recency? Inheriting *βo:xʒɑ gives the Herta-Olaic four new, fresh, recent; young, unfamiliar; restricting them gives new, recent, the narrower and more conservative reading. *woge has no parent entry and must be stated outright. beza is a borderline sixth case: its correction above is mechanical, but the entry records only the first of its parent's four senses and so does not itself say whether Besoh "new" is recency or novelty — the operator may wish to pin it with the bare rows.

dull

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*mjaəni	dull	blunt, not sharp / (by extension) slow-witted — the word "dull" never appears
Ancient Hertic	ιαμνιε	dull	"dull" — sole line
Gaillean	maennia	dull	"dull" — sole line
Middle Airananwe	mainya	dull	"dull" — sole line
Airananwe	maignay	dull	"dull" — sole line
Proto-Njeshan	*komlit	dull	"dull" — sole line
Proto-Andusian	*mɛᵐb	dull	dull, blunt / (by extension) slow, stupid
Proto-North-Andusian	*meᵐb	dull	dull, blunt / (by extension) slow, stupid
Proto-South-Andusian	*mɛm	dull	dull, blunt / (by extension) slow, stupid
Proto-Unterian	*mɛmu	dull	dull, blunt / (by extension) slow, stupid
Besoh	mem	dull	dull, blunt / (by extension) slow, stupid

Answerable from the entries. Set *mjaəni to blunt; slow-witted and the five Andusian entries *mɛᵐb, *meᵐb, *mɛm, *mɛmu and mem to blunt (of an edge); slow, stupid — dropping "dull" from "dull, blunt" is itself the disambiguation, since blunt is the unambiguous half. *mjaəni is the sharpest case in the section: its entry never uses the word "dull" at all, so its gloss is not merely underspecified but wrong. No entry in either family records the colour or light sense of English dull, nor the "boring" sense; only the blade and the wit are attested, and the English headword blunt would describe this whole set better than dull does — worth considering as a rename of the gloss key, which would disturb no entry, since — unlike big, long and new — this gloss has no English lexeme of its own in the Dictionary.

Needs a ruling. Do the bare ιαμνιε, maennia, mainya, maignay and *komlit carry the extended sense of wit as well as the physical edge? Inheriting *mjaəni gives the Herta-Olaic four blunt; slow-witted; restricting them to the parent's first line gives blunt alone, with the figurative extension treated as a Proto-Herta-Olaic development the daughters did not keep. *komlit has no parent entry; its own line is the bare word "dull", so if it is meant to reach colour or tedium — senses nothing else in the corpus attests — that is information only the operator can supply.

thick

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*k͡xuəʃ	thick	one line: "thick, dense, stout"
Ancient Hertic	γοζε	thick	"thick" — sole line
Gaillean	ghouca	thick	"thick" — sole line
Middle Airananwe	ghuuca	thick	"thick" — sole line
Airananwe	houchay	thick	"thick" — sole line
Proto-Andusian	*dzof	thick	thick / dense (of forest, of fog) / stout
Proto-North-Andusian	*dzof	thick	thick / dense (of forest, of fog) / stout
Proto-South-Andusian	*dzof	thick	thick / dense (of forest, of fog) / stout
Proto-Unterian	*zofi	thick	thick / dense (of forest, of fog) / stout
Besoh	jof	thick	thick / dense (of forest, of fog) / stout

Answerable from the entries. Set *k͡xuəʃ to thick (in cross-section); dense; stout and the five Andusian entries *dzof, *dzof, *dzof, *zofi and jof to thick (in cross-section); dense (of forest, fog); stout. Three of those share both the spelling dzof and the page Dictionary:dzof, where they sit as three language sections, so the gloss is again all that separates them in a list.

Needs a ruling. Do the bare γοζε, ghouca, ghuuca and houchay cover density and bodily stoutness, or thickness in cross-section alone? Inheriting *k͡xuəʃ gives thick, dense, stout; restricting gives thick (in cross-section), and a fog or a forest would then need another word in all four languages. There is no Njeshan entry for this gloss. This should be decided together with thin.

thin

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*g͡ɣɛnki	thin	one line: "thin, slender, sparse"
Ancient Hertic	ενδεκε	thin	"thin" — sole line
Gaillean	gheka	thin	"thin" — sole line
Middle Airananwe	ghenke	thin	"thin" — sole line
Airananwe	ghenney	thin	"thin" — sole line
Proto-Andusian	*rubʰ	thin	thin / lean, gaunt / fine, slender
Proto-North-Andusian	*rub	thin	thin / lean, gaunt / fine, slender
Proto-South-Andusian	*ruː	thin	thin / lean, gaunt / fine, slender
Proto-Unterian	*ruwi	thin	thin / lean, gaunt / fine, slender
Besoh	rü	thin	thin / lean, gaunt / fine, slender

Answerable from the entries. Set *g͡ɣɛnki to thin (in cross-section); slender; sparse and the five Andusian entries *rubʰ, *rub, *ruː, *ruwi and rü to thin (in cross-section); lean (of a body); fine, slender. Unlike thick, no two forms here share a spelling. There is no Njeshan entry for this gloss.

Needs a ruling. Do the bare ενδεκε, gheka, ghenke and ghenney extend to bodily leanness and to sparseness? Inheriting *g͡ɣɛnki gives thin, slender, sparse; restricting gives thin (in cross-section). Note the asymmetry a split decision would leave: Proto-Herta-Olaic records sparseness for thin and density for thick, a reciprocal pair, so reading one broadly and the other narrowly would break it.

heavy

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*xətʃoʊ	heavy	heavy, weighty / (by extension) grave and burdensome
Ancient Hertic	ετζωιε	heavy	"heavy" — sole line
Gaillean	hettowa	heavy	"heavy" — sole line
Middle Airananwe	hyettoe	heavy	"heavy" — sole line
Airananwe	yettoy	heavy	"heavy" — sole line
Proto-Andusian	*ᵑgia	heavy	heavy, weighty / (by extension) burdensome / grave, important / (figurative) solemn / oppressive
Proto-North-Andusian	*ᵑgia	heavy	heavy, weighty / (by extension) burdensome
Proto-South-Andusian	*gia	heavy	heavy, weighty / (by extension) burdensome
Proto-Unterian	*ŋia	heavy	heavy, weighty / (by extension) burdensome
Besoh	gia	heavy	heavy, weighty / (by extension) burdensome / (by extension) mighty, strong

Answerable from the entries. Set *xətʃoʊ to heavy (in weight); grave, burdensome; *ᵑgia to heavy (in weight); burdensome; grave, solemn, oppressive; *ᵑgia, *gia and *ŋia to heavy (in weight); burdensome; gia to heavy (in weight); burdensome; mighty, strong. Two spelling clashes ride on this gloss and they run in opposite directions: *ᵑgia and *ᵑgia share the page Dictionary:ᵑgia, where the parent carries three senses the daughter has dropped; *gia and gia share the page Dictionary:gia, where the Besoh reflex has instead added a sense — "mighty, strong" — that its parent lacks. In both cases the gloss is the only field that will show which is which.

Needs a ruling. Do the bare ετζωιε, hettowa, hyettoe and yettoy carry the figurative sense of gravity and burden alongside physical weight? Inheriting *xətʃoʊ gives heavy (in weight); grave, burdensome; restricting gives heavy (in weight), which would leave Herta-Olaic the only branch with no figurative extension of this root, against an Andusian set in which every reflex has one. There is no Njeshan entry for this gloss.

full

Language	Word	Current gloss	What the entry's own definitions say
Proto-Herta-Olaic	*ktu:lhʃa	full	full, filled, complete / (by extension) satisfied
Ancient Hertic	τυχασε	full	"full" — sole line
Gaillean	kullenca	full	"full" — sole line
Middle Airananwe	kuulence	full	"full" — sole line
Airananwe	coulsey	full	"full" — sole line
Proto-Andusian	*dʰixwa	full	full, filled / sated / (by extension) whole, complete / (figurative) satisfied, content
Proto-North-Andusian	*dexwa	full	full, filled / sated / (by extension) whole, complete
Proto-South-Andusian	*hixwa	full	full, filled / sated / (by extension) whole, complete
Proto-Unterian	*riwa	full	full, filled / sated / (by extension) whole, complete
Besoh	hiĥva	full	full, filled / sated / (by extension) whole, complete

Answerable from the entries. Set *ktu:lhʃa to full, filled, complete; satisfied; *dʰixwa to full, filled; sated; whole, complete; satisfied, content; and *dexwa, *hixwa, *riwa and hiĥva to full, filled; sated; whole, complete. This gloss runs on three axes rather than two — a filled container, a sated person, a whole quantity — and the two families slice them differently: Andusian keeps all three apart and adds contentment as a fourth in the proto, whereas Proto-Herta-Olaic merges "filled" and "complete" on one line and treats satisfaction as the extension.

Needs a ruling. Do the bare τυχασε, kullenca, kuulence and coulsey cover satiety and completeness as well as a filled container? Inheriting *ktu:lhʃa gives full, filled, complete; satisfied; restricting gives full, filled, and the branch then needs separate roots for "sated" and for "whole", neither of which is yet coined in any of the four languages. There is no Njeshan entry for this gloss.

Distinctions English lacks

we

Language	Word	Current gloss	What the entry's own definition says
Proto-Herta-Olaic	*g͡ɣahəls	we	“first person plural” — person and number only, nothing on clusivity
Ancient Hertic	γαισ	we	“we”
Gaillean	ghale	we	“we”
Middle Airananwe	ghaal	we	“we”
Airananwe	ghá	we	“we”
Proto-Njeshan	*neβon	we	“we”
Proto-Andusian	*guᵐbɛ	we (inclusive)	“we two, you and I”; “we (inclusive)”. Etymology: *gum (“you”) + *bɛ (“I”)
Proto-Andusian	*qodʰ	we	“we”, plus a usage note: whether it included the hearer “is not recoverable”, and beside *guᵐbɛ (“you and I”) it “is conventionally read as exclusive”
Proto-North-Andusian	*ʔod	we	“we”, from *qodʰ; the same stage carries *guᵐbe “we two, you and I / we (inclusive)”
Proto-South-Andusian	*qoː	we	“we”, from *qodʰ; beside *gumɛ “we two, you and I / we (inclusive)”
Proto-Unterian	*ʔori	we	“we”, from *qodʰ; beside *gumɛ “we two, you and I / we (inclusive)”
Besoh	qö	we	“we”, from *qoː; beside gume “we two, you and I / we (inclusive)”

Answerable from the entries. Eleven lemmas carry the bare gloss we, but the Andusian branch is not among the unmarked: Dictionary:qodʰ states in its own usage notes that beside its partner it is conventionally read as exclusive, so *qodʰ should read we (exclusive). The contrast survives to every daughter stage — each of *ʔod, *qoː, *ʔori and qö descends from qodʰ and stands beside an inherited reflex of *guᵐbɛ carrying the inclusive definition word for word — so those four should read we (exclusive) as well; if the operator prefers a gloss to rest only on what its own page says, that inference is the one part to withhold, and qodʰ alone changes. One inconsistency travels with the fix: *guᵐbɛ is glossed “we (inclusive)” while its four descendants *guᵐbe, *gumɛ, *gumɛ and gume are glossed “we two”, though all six carry identical definition lines. The clusivity form is the one that pairs with the fix above and matches the family's other pronoun pair, so the four should take we (inclusive), the dual reading staying on the first definition line.

Needs a ruling. Do the Herta-Olaic pronouns and Proto-Njeshan *neβon distinguish inclusive from exclusive at all? If they do not, clusivity is an Andusian feature, the six glosses stay “we”, and the “first person plural” of *g͡ɣahəls is the correct reading for the whole branch; if they do, each of *g͡ɣahəls, γαισ, ghale, ghaal, ghá and *neβon becomes “we (exclusive)” and six partner pronouns must be coined into a paradigm otherwise reconstructed unbroken from proto to Airananwe. Nothing on the six pages bears on it: the Herta-Olaic pronoun system is documented for number only (*ixaβ plural against *nwijɔ singular, ruling 8 above), and the Njeshan corpus has no paired pronoun of any kind, so its silence is especially weak evidence.

know

Language	Word	Current gloss	What the entry's own definition says
Proto-Herta-Olaic	*hɑtʲwɑ	know	“know, be familiar with, recognise”, with a related-terms line “Compare: *ɲijupt (see)”
Ancient Hertic	χαδουισ	know	“know”
Gaillean	athwei	know	“know”
Middle Airananwe	oothuee	know	“know”
Airananwe	authé	know	“know”
Proto-Njeshan	*tarwil	know	“know”
Proto-Andusian	*emi	know	“know”; “recognize”; “know how, be able”; (figurative) “to be intimate with”
Proto-North-Andusian	*emi	know	“know”; “acknowledge”
Proto-South-Andusian	*emi	know	“know”; “recognize”; “know how, be able”
Proto-Unterian	*emi	know	“know”; “know how, be able”
Besoh	emi	know	“know”; “recognize”; “know how, be able” — beside keśmi “to know how, be able; manage, succeed”, a compound of keś (“do”) and emi

Answerable from the entries. Six of the eleven already state their range and only the gloss lags behind. *hɑtʲwɑ should read know, be familiar with — its definition names familiarity and recognition, which is the acquaintance sense, not the propositional one. The Andusian series should follow its own lines: *emi and *emi know, recognise; know how; *emi know; know how; *emi know, acknowledge. The ability sense in particular has no business being invisible: a reader of a translations box would otherwise never learn that these words also render “know how to”. Besoh is the exception — emi should take know, recognise only, because the ability sense is separately lexicalised there as keśmi, whose gloss is already “know how”; whether the two Besoh lemmas should both hold that sense is a collision of its own, invisible to the check that built this page because their gloss fields differ. Note also that the entries spell the recognition sense two ways — recognise on *hɑtʲwɑ, recognize on the four Andusian pages — and the glosses above regularise on the first.

Needs a ruling. For the four reflexes of *hɑtʲwɑ and for *tarwil, does one verb cover knowing a fact, being acquainted with a person, and knowing how — or does the language split them? If it is one verb, the glosses become “know (a fact or a person)” and the branch carries the proto's range forward unremarked; if it splits, χαδουισ, athwei, oothuee and authé narrow to propositional knowing — a change of sense against their own etymology, since the proto they inherit from is explicitly the familiarity verb — and each language then needs a new acquaintance verb. Nothing in the corpus is glossed “learn”, “understand” or “remember” in any language, and the only neighbouring verb anywhere is the Besoh compound keśmi.

say

Language	Word	Current gloss	What the entry's own definition says
Proto-Herta-Olaic	*nwa:ɛɲ	say	“say, speak, utter”, with “Compare: *upɔtʃi (name)”
Ancient Hertic	νουιω	say	“say”; the language has no verb glossed “speak”
Gaillean	naenei	say	“say”; beside dinei “speak”
Middle Airananwe	nayanee	say	“say”; beside dienee “speak”
Airananwe	naié	say	“say”; beside tiengé “speak”
Proto-Njeshan	*djuka	say	“say”
Proto-Andusian	*gɔɛ	say	“say, speak, tell”; (by extension) “call, name”; (figurative) “to promise”; “to command”
Proto-North-Andusian	*goe	say	“say”; “pronounce, decree”
Proto-South-Andusian	*gɔː	say	“say, speak”; “tell”
Proto-Unterian	*gɔɛ	say	“say, speak”; “promise”
Besoh	gö	say	“say, speak”; “tell” — etymology 2 of the page; etymology 1 is an unrelated noun “louse”, whose senses are not this word's

Answerable from the entries. The whole Andusian series and the Herta-Olaic proto already say how wide they are: *nwa:ɛɲ → say, speak, utter; *gɔɛ → say, speak, tell; *gɔː and gö → say, speak, tell; *gɔɛ → say, speak; *goe → say, pronounce. Three further glosses need no change and should not be re-flagged: naenei, nayanee and naié each stand against a dedicated speak-verb in the same language, so “say” is already delimited on the side that matters.

Needs a ruling. Do νουιω and *djuka cover speech generally, as their protos do, or only the uttering of words? If generally, both become “say, speak” and Ancient Hertic keeps the undivided proto verb; if narrowly, both stay “say” and each language needs a speak-verb — which would make Gaillean dinei the model for a lexeme Ancient Hertic never acquired on the same branch, though dinei carries {{etym-missing}} and so is not itself inherited from *nwa:ɛɲ. The corpus is thin here and the absence proves little: outside Airananwe it holds no word for “language” or “word” either.

smell

Language	Word	Current gloss	What the entry's own definition says
Proto-Herta-Olaic	*b͡βɪɲo:	smell	“smell, perceive by nose”; “give off a smell” — both directions, explicitly
Ancient Hertic	φιουνισ	smell	“smell”
Gaillean	bhinwei	smell	“smell”
Middle Airananwe	bhinuee	smell	“smell”
Airananwe	bhynué	smell	“smell”
Proto-Andusian	*femug	smell	“smell”; “sniff”; (by extension) “track by scent”
Proto-North-Andusian	*femug	smell	“smell”; “sniff”; (by extension) “track by scent”
Proto-South-Andusian	*femuʔ	smell	“smell”; “sniff”; (by extension) “track by scent”
Proto-Unterian	*femugi	smell	“smell”; “sniff”; (by extension) “track by scent”
Besoh	femu'	smell	“smell”; “sniff”; (by extension) “track by scent”

Answerable from the entries. *b͡βɪɲo: spells out both directions and its gloss should say so: smell (perceive or give off an odour). The five Andusian words are agentive perception throughout — every sense they record is sniffing or tracking by scent, and no emission sense appears anywhere in the branch — so *femug, *femug, *femuʔ, *femugi and femu' should read smell (perceive an odour), leaving “sniff” and “track by scent” to the definition lines.

Needs a ruling. Do the four reflexes of *b͡βɪɲo: keep the proto's second sense, “give off a smell”? If they do, φιουνισ, bhinwei, bhinuee and bhynué take “smell (perceive or give off an odour)” and the alternation becomes a stable branch feature; if they do not, they take “smell (perceive an odour)” and four languages are left with no way to say that something stinks, since no lexeme in the corpus is glossed “stink”, “reek” or “odour” and there is no noun for a smell in any language — so the second half of that ruling is a commission, not just a gloss.

come

Language	Word	Current gloss	What the entry's own definition says
Proto-Herta-Olaic	*k͡xiɛɹt	come	“come, approach, arrive — motion toward the speaker”
Ancient Hertic	γιτσισ	come	“come”
Gaillean	ghirtei	come	“come”
Middle Airananwe	ghirtee	come	“come”
Airananwe	ghité	come	“come”
Proto-Njeshan	*kavilm	come	“come”
Proto-Andusian	*kɛ	come	“come, arrive”; (by extension) “become, come to be”; (figurative) “to happen, come to pass”; antonym *dʰes “go”
Proto-North-Andusian	*ke	come	“come, arrive”; “visit”
Proto-South-Andusian	*kɛ	come	“come”; “happen, befall” — etymology 2 of the page; etymology 1 is a conjunction, “because, for / from, out of”
Proto-Unterian	*kɛ	come	“come”; “become”
Besoh	ke	come	“come”; “happen, befall” — likewise etymology 2, beside the same conjunction

Answerable from the entries. Six glosses hide a sense their own page records, and none of the six fixes touches the deictic question below. *k͡xiɛɹt → come, approach, arrive; *kɛ → come, arrive; become; *ke → come, arrive; *kɛ → come; become; *kɛ and ke → come; happen. The copular and eventive senses are worth the space — they are what a translations box would most easily send a reader wrong on — where the motion senses “visit” and “arrive” can stay on the definition lines.

Needs a ruling. Does a come-verb cover approach short of arrival as well as arrival itself, and is its motion reckoned toward the speaker as *k͡xiɛɹt alone says? Only the Herta-Olaic proto records either point; *kɛ names arrival but not approach, and the other nine say only “come”. If the proto's reading holds generally, the ten inherit “come, approach, arrive” and the deixis is worth putting in the gloss as “(toward the speaker)”, since a translations box is read without the entry beside it; if approach is excluded, every one of those languages needs an approach-verb and the corpus has none — nothing anywhere is glossed “approach”, “arrive”, “depart” or “leave”, the stock of path verbs running to the walk-series (*cɔki “walk, go / travel on foot”, *uəɲu), the Andusian go-series, and one Besoh derivative, lavñevä “return, turn back”.

go

Language	Word	Current gloss	What the entry's own definition says
Proto-Andusian	*dʰes	go	“go, depart”; “set out”; “leave”; (figurative) “to die (euphemistic)”; antonym *kɛ “come”
Proto-North-Andusian	*dɛs	go	“go”; “travel”
Proto-South-Andusian	*hes	go	“go, depart”; “set out”; “leave”
Proto-Unterian	*resi	go	“go, depart”; “die (of elders)”
Besoh	hes	go	“go, depart”; “set out”; “leave”

Answerable from the entries. This one is settled twice over. The English lemma Dictionary:go — the only one of the six glosses in this section to have one — defines go as “To move away from where one is; to depart” and, euphemistically, “To die”, and its translations box carries the sense gloss “to move away from where one is; to depart” with all five Andusian words under it; each of the five points back at that lemma from its own first definition line. Four of them then say “depart / set out / leave” in their own words, and the branch lexicalises the contrast independently: *dʰes is departure, *cɔki “walk, go / travel on foot” is locomotion. The glosses should read go, depart for *dʰes, *hes, hes and *resi, and go, depart; travel for *dɛs, the one member whose own definition omits departure while the translations box still files it under that sense. The euphemistic death sense on *dʰes and *resi mirrors sense 2 of the English lemma and needs no room in the gloss. No word outside Andusian is glossed “go”, so nothing here touches the other two families, and row 50's own verdict — that the Andusian go family is ruled as depart — is already on the page; only the gloss field is outstanding.