Why was Spanish the only Romance language to lose the initial "F" in Latin words?
Going through the History of the Spanish language article in Wikipedia, I read the section Latin f- to Spanish h- to null some interesting insight:
F was almost always initial in Latin words, and in Spanish most of them passed through a stage in which the consonant eventually developed to [h] and then was lost phonologically.
And also:
The change from /f/ to /h/ occurred in the Romance speech of Old Castile and Gascon, but nowhere else nearby. Since both areas were historically bilingual with Basque, and Basque once had [h] but no [f], it is often suggested that the change was caused by Basque influence. However, this is contested by many linguists.
So it is interesting to see that this is indeed the case just for Spanish. See some examples:
| Latin | English | Spanish | Catalan | Portuguese | Italian |
|---|---|---|---|---|---|
| filum | thread | hilo | fil | fio | filo |
| filius | son | hijo | fill | filho | figlio |
| furnus | oven | horno | forn | forno | forno |
As indicated in the quote, this is just specific to Spanish. But, what other theories are there to explain this? I couldn't find them. Does the usage of Arabic throughout the Iberian Peninsula during the VIII to XVI centuries have any influence on this different evolution in contrast with the other Romance languages?
0 answers
8 comments
Arabic has initial "f" and Spanish actually retains it easily, although it was lost in some early loanwords as well. https://en.wikipedia.org/wiki/Arabic_language_influence_on_the_Spanish_language#F,_G — Jirka Hanika about 2 months ago
According to the Wikipedia this is an unsolved mistery. — . . about 2 months ago
"En resumen, se puede decir que nadie ha analizado satisfactoriamente la complejidad del fenómeno [...]". — . . about 2 months ago
The same sound change happened in Japanese (partially) — user8078 about 2 months ago
Not just in Japanese [not affirming myself], also in central Sardinian and maybe elsewhere in Romance. If anyone has any examples of Latin "f" which is not morpheme-initial, that might allow additional research into whether the particular developments appear the same or only partly the same, and thus likely related or likely independent. — Jirka Hanika about 2 months ago