I think you may be being misled by the transcription in CORDE's database. Accents were not generally used in Spanish until the 16th century:
In general, the accent mark was not employed in manuscripts in Spanish or Spanish books before 1550. It was only after the Italian and French models had been created that the widespread use of accent marks began to be practiced in Spanish. Between 1475 (the approximate date of the introduction of printing into Spain) and 1566, the use of accent marks was negligible
So it seems likely that the editions of these works CORDE has scanned are from a later copy than the early handwritten manuscripts.
1. Orthography: Coexistence of sandío and sandio given the common pronunciation /san'dio/
Spanish's history of accents prior to the 19th century is a long, complicated, and often contradictory affair. Limiting ourselves to discussion of llana words ending -io/-ío, the ambiguities with stress are thus:
When the last syllable of a noun or a verb contained a non-Latin diphthong (the
only Classical Latin diphthongs had been oe and ae, and many new diphthongs had been created in Vulgar Latin and in Spanish itself), a new problem developed: what about a word like diferencia, which by Classical Latin rules
could be accented on the final i, because it constituted a separate syllable? Should it be spelled diferéncia or should the modern language accept the notion of the diphthongs that had developed as early as Vulgar Latin times?
We note that in the "Discurso proemial de la orthographia" (Discurso) of the first Diccionario (1726), "the first statement on correctness of Spanish spelling by an official governing body" they did not address this question at all:
It is interesting to note that, even with all the fuss concerning long vowels, very basic questions were not addressed, such as the issue of how to show stress on words like diferencia and mengua as opposed to vacío and evalúa... The authors of the Discurso simply ignored the issue altogether.
This ambiguity can be seen in the frequent inconsistency among even the writers of the Discurso itself in treatment of accents on such words:
The treatment of paroxytones is quite chaotic. We find calidades and calidádes on the very same page just a few lines apart. Not so close are comúnes and divisiones, solía and dia, principáles and universales. In each of these pairs of words the identical pattern of vowels and consonants occurs with the identical pattern of accentuation; but half of them have accents and half of them do not. The words orígen and léngua always have a written accent. Words ending in diphthongs -io or -ia usually carried an accent on the preceding stressed vowel, as in necessários, princípio, próprios, and observáncia, but we also find ignorancia and diferencia.
So, given all this, it is unsurprising that there is a mix of orthographies for sandío/sandio in the CORDE's register:
Indeed, looking at the NTLLE, we see that even within dictionary headers, usage is inconsistent year on year (even within the same publishing bodies) until the second half of the 19th century:
- sandio 1495 NEBRIJA
- ...
- sandio 1611 COVARRUBIAS
- sandío 1611 ROSAL
- sandio 1617 MINSHEU
- sandio 1620 FRANCIOSINI
- sandio 1670 MEZ DE BRAIDENBACH
- sándio 1705 SOBRINO
- sandío 1706 STEVENS
- sandio 1721 BLUTEAU
- sandio 1739 ACADEMIA AUTORIDADES (S-Z)
- sándio 1780 ACADEMIA USUAL
- sandio 1783 ACADEMIA USUAL
- ...
- sandio 1803 ACADEMIA USUAL
- sandío 1817 ACADEMIA USUAL
- sándio 1822 ACADEMIA USUAL
- sandio 1825 NÚÑEZ DE TABOADA
- ...
- sandio 1846 SALVÁ
- sándio 1852 ACADEMIA USUAL
- sandio 1853 DOMÍNGUEZ
- sándio 1855 GASPAR Y ROIG (G-Z)
- sandio 1869 ACADEMIA USUAL
- ...
- sandio 1992 ACADEMIA USUAL
2. Pronunciation: /san'dio/ → /'sandjo/
This still leaves us with the question of, assuming it was originally pronounced /san'dio/, how it came to be pronounced /'sandjo/ (and thus why the spelling sandío fell out of use).
As per the NTLLE search, it seems the /'sandjo/ pronunciation existed since at least the start of the 18th century (and possibly much earlier). Assuming this pronunciation was a later development, there is ample precedent for such a shift. Consider the following:
b). Diptongo (Hiato), (62):
carau (caraú), balaustre (balaústre), bronquiolo (bronquíolo),
acedia (acedía), ademprio (ademprío), agrafia (agrafía), elefancia (elefancía),
aeromancia (aeromancía), alectomancia (alectomancía), bibliomancia (bibliomancía), capnomancia (capnomancía), cartomancia (cartomancía), catoptromancia (catoptromancía), ceraunomancia (ceraunomancía), ceromancia (ceromancía), demonomancia (demonomancía), espatulomancia (espatulomancía), exoftalmia (exoftalmía), geomancia (geomancía), heteromancia (heteromancía), hidromancia (hidromancía), lecanomancia (lecanomancía), necromancia (necromancía), oniromancia (oniromancía), onomancia (onomancía), ornitomancia (ornitomancía), piromancia (piromancía), quiromancia (quiromancía), uromancia (uromancía),
oftalmia (oftalmía), xeroftalmia (xeroftalmía),
-plastia (-plastía (Amér)), hernioplastia (hernioplastía),
cuadriplejia (cuadriplejía), hemiplejia (hemiplejía), paraplejia (paraplejía), tetraplejia (tetraplejía),
-scopia (-scopía (Amér)), endoscopia (endoscopía (Amér)), gastroscopia (gastroscopía (Amér)), microscopia (microscopía),
olimpiada (olimpíada), paralimpiada (paralimpíada), parolimpiada (parolimpíada),
amoniaco (amoníaco), armoniaco (armoníaco (desus)), austriaco (austríaco), bosniaco (bosníaco), cardiaca (cardíaca (Amér)), cordiaco (cordíaco (desus)) egipciaco (egipcíaco), isiaco (isíaco), lipemaniaco (lipemaníaco), monomaniaco (monomaníaco), olimpiaco (olimpíaco (desus)), peloponesiaco (peloponesíaco), simoniaco (simoníaco), siriaco (siríaco), zodiaco (zodíaco).
etc
Sources:
