Invitational round | 10 points | 80.00% | Problem statement | Official solution | Tags: PhonologyHistorical
In R1:
All this is just to get familiar with the rule notation and mechanics. Now onto real business.
For R2, we just look at the last rule and see what language contains the feature that was generated by that rule. Language A's last change is (17) → ei, so it should contain some "ei" bigrams. Only Portuguese has them. Language C's last change is (23) → dd, so it should contain some "dd" bigrams. Only Sardinian has them. That leaves B as Italian.
A: Portuguese
B: Italian
C: Sardinian
For R3, we can do it part-by-part. First consider Latin to Vulgar Latin. Take all Latin words and apply all the rules we already know.
(I will abbreviate [+Vow] and [+Const] as "V" and "C", but expand them in the final solution.)
| Latin | h→∅ | m→∅ / _# | (1)→∅ / _# | (2)→(3) / V_V | (4)→(5) / _t | ∅→i / #_sC |
|---|---|---|---|---|---|---|
| focum | focum | focu | focu | focu | ||
| bonum | bonum | bonu | bonu | bonu | ||
| ruptum | ruptum | ruptu | ruttu | ruttu | ||
| caballum | caballum | caballu | cavallu | cavallu | ||
| sonāre | sonāre | sonāre | sonāre | sonāre | ||
| pulvera | pulvera | pulvera | pulvera | pulvera | ||
| nivis | nivis | nivis | nivi | nivi | ||
| penna | penna | penna | penna | penna | ||
| sinum | sinum | sinu | sinu | sinu | ||
| fricāre | fricāre | fricāre | fricāre | fricāre | ||
| adoptāre | adoptāre | adoptāre | adottāre | adottāre | ||
| virdem | virdem | virde | virde | virde | ||
| arēna | arēna | arēna | arēna | arēna | ||
| cibus | cibus | cibus | civu | civu | ||
| habere | abere | abere | avere | avere | ||
| vanum | vanum | vanu | vanu | vanu | ||
| siccum | siccum | siccu | siccu | siccu | ||
| pilum | pilum | pilu | pilu | pilu | ||
| mittere | mittere | mittere | mittere | mittere | ||
| vīlla | vīlla | vīlla | vīlla | vīlla | ||
| lumbum | lumbum | lumbu | lumbu | lumbu | ||
| statum | statum | statu | statu | istatu | ||
| spatha | spata | spata | spata | ispata | ||
| pira | pira | pira | pira | pira | ||
| pullum | pullum | pullu | pullu | pullu | ||
| castellum | castellum | castellu | castellu | castellu | ||
| marītum | marītum | marītu | marītu | marītu | ||
| lūna | lūna | lūna | lūna | lūna |
Next, from Vulgar Latin to Italian (because the chain is shorter). Here, we can first back-deduce from e→∅ / #_sC. There are only two words that start with sC; and their Vulgar Latin form both start with "i", so we assume that this "i" changed into "e" which eventually gets dropped. The first rule that says change something into "e" should probably do that job. Indeed, all other "i" changed into "e" as well, such as siccu→secco, not just word initial. So (6) = i.
| VL | i→e | (7)→(8) | V[+L]→V[-L] | (18)→(19) / (20) | e→∅ / #_sC |
|---|---|---|---|---|---|
| focu | focu | fuoco | fuoco | ||
| bonu | bonu | buono | buono | ||
| ruttu | ruttu | rotto | rotto | ||
| cavallu | cavallu | cavallo | cavallo | ||
| sonāre | sonāre | suonare | suonare | ||
| pulvera | pulvera | polvera | polvera | ||
| nivi | neve | neve | neve | ||
| penna | penna | penna | penna | ||
| sinu | senu | seno | seno | ||
| fricāre | frecāre | frecare | frecare | ||
| adottāre | adottāre | adottare | adottare | ||
| virde | verde | verde | verde | ||
| arēna | arēna | arena | arena | ||
| avere | avere | avere | avere | ||
| vanu | vanu | vano | vano | ||
| siccu | seccu | secco | secco | ||
| pilu | pelu | pelo | pelo | ||
| mittere | mettere | mettere | mettere | ||
| vīlla | vīlla | villa | villa | ||
| lumbu | lumbu | lombo | lombo | ||
| istatu | estatu | estato | stato | ||
| ispata | espata | espata | spata | ||
| pira | pera | pera | pera | ||
| pullu | pullu | pollo | pollo | ||
| castellu | castellu | castello | castello | ||
| marītu | marītu | marito | marito | ||
| lūna | lūna | luna | luna |
Other than long vowel to short vowel, there's another universal sound change: all "u" become "o", except:
Because we have "u" sounds that survive (especially "luna"), the rule u→o should come very early, so that later "u" can be reinserted by other rules. Therefore, (7) = u, (8) = o. Once we have this, we can apply the V[+L]→V[-L] rule to get the before/after of the (18)→(19) / (20) change.
| Before | After | Before | After |
|---|---|---|---|
| foco | fuoco | vano | vano |
| bono | buono | secco | secco |
| rotto | rotto | pelo | pelo |
| cavallo | cavallo | mettere | mettere |
| sonare | suonare | villa | villa |
| polvera | polvera | lombo | lombo |
| neve | neve | estato | estato |
| penna | penna | espata | espata |
| seno | seno | pera | pera |
| frecare | frecare | pollo | pollo |
| adottare | adottare | castello | castello |
| verde | verde | marito | marito |
| arena | arena | luna | luna |
| avere | avere |
Therefore, this rule only changes "o" to "uo" in "foco", "bono", "sonare", but not "polvera", "lombo", or all word-final "o"s. The rule o→uo / _CV will do this job, although there are many other rules that could also explain the data at hand.
Now we know the three rewrite rules that lead to Western Romance, we can use that as the starting point for deducing Portuguese. The e→∅ / (16) rule deletes some "e"s; the only "e"s that are deleted are the word-final ones such as "frecare" and "adottare". However, the "e" in "verde" or "neve" are not deleted; only those after "r" are. So Therefore (16) = r_ (we don't have cases where "re" is not at the end, so I don't know if we also need # at the end).
| WR | (9)→(10) / (11) | (12)→(13) / _V | (14)→(15) | e→∅ / (16) | (17)→ei |
|---|---|---|---|---|---|
| foco | fogo | fogo | fogo | ||
| bono | bõo | bõo | bõo | ||
| rotto | roto | roto | roto | ||
| cavallo | cavalo | cavalo | cavalo | ||
| sonare | sõar | sõar | sõar | ||
| neve | neve | neve | neve | ||
| penna | pena | pena | pena | ||
| seno | s(17)o | s(17)o | seio | ||
| frecare | fregare | fregar | fregar | ||
| adottare | adotare | adotar | adotar | ||
| verde | verde | verde | verde | ||
| arena | ar(17)a | ar(17)a | areia | ||
| cevo | cevo | cevo | cevo | ||
| avere | avere | aver | aver | ||
| vano | vão | vão | vão | ||
| secco | seco | seco | seco | ||
| pelo | pelo | pelo | pelo | ||
| mettere | metere | meter | meter | ||
| villa | vila | vila | vila | ||
| lombo | lombo | lombo | lombo | ||
| estato | estado | estado | estado | ||
| espata | espada | espada | espada | ||
| pera | pera | pera | pera | ||
| pollo | polo | polo | polo | ||
| castello | castelo | castelo | castelo | ||
| marito | marido | marido | marido | ||
| luna | lũa | lũa | lũa |
Skimming the data, there are three changes:
We just need to devise the environment for each change.
Finally, we still need to figure out what (17) is in (17)→ei. Applying all previous rules, we have seno→sẽo, arena→arẽa. Therefore (17) = ẽ. Indeed, we have no instance of ẽ in the final data.
Lastly, from Vulgar Latin to Sardinian.
| VL | V[+L]→V[-L] | (21)→(22) | (23)→dd |
|---|---|---|---|
| focu | focu | focu | focu |
| bonu | bonu | bonu | bonu |
| ruttu | ruttu | ruttu | ruttu |
| cavallu | cavallu | caba(23)u | cabaddu |
| sonāre | sonare | sonare | sonare |
| pulvera | pulvera | pulbera | pulbera |
| nivi | nivi | nibi | nibi |
| sinu | sinu | sinu | sinu |
| fricāre | fricare | fricare | fricare |
| virde | virde | birde | birde |
| arēna | arena | arena | arena |
| civu | civu | cibu | cibu |
| avere | avere | abere | abere |
| siccu | siccu | siccu | siccu |
| pilu | pilu | pilu | pilu |
| mittere | mittere | mittere | mittere |
| vīlla | villa | bi(23)a | bidda |
| istatu | istatu | istatu | istatu |
| ispata | ispata | ispata | ispata |
| pira | pira | pira | pira |
| pullu | pullu | pu(23)u | puddu |
| castellu | castellu | caste(23)u | casteddu |
| lūna | luna | luna | luna |
The answer is apparent: (21) = v, (22) = b, (23) = ll.
To recap:
(1) s
(2) b
(3) v
(4) p
(5) t
(6) i
(7) u
(8) o
(9) [+Stop -Voiced]
(10) [+Stop +Voiced]
(11) [+Vow]_[+Vow]
(12) [+Vow]n
(13) [+Vow +Nasal]
(14) [+Const]{2}
(15) [+Const]{1}
(16) r_
(17) ẽ
(18) o
(19) uo
(20) _[+Const][+Vow]
(21) v
(22) b
(23) ll
Finally, rewrite from Latin to Western Romance:
| Rule | baculum | vīta | findāre |
|---|---|---|---|
| h→∅ | — | — | — |
| m→∅ / _# | baculu | — | — |
| s→∅ / _# | — | — | — |
| b→v / V_V | — | — | — |
| p→t / _t | — | — | — |
| ∅→i / #_sC | — | — | — |
| i→e | — | — | fendāre |
| u→o | bacolo | — | — |
| V[+L]→V[-L] | — | vita | fendare |
Therefore, they become: bacolo, vita, fendare.