7

At some point during the evolution of Spanish, several initial [f] became silent (this is represented with an h in modern Spanish). This explains words such as hacer, harina, herir and many more. However, there are many words in modern Spanish which obviously come from Latin and have retained the initial [f]: forma, fácil, fatal, fuerte... (the last one has diphthongized, but clearly comes from fortis, so one could wonder why do we have horma but not hortis/horte).

What explains all these irregularities? At what point was this sound change operating and why did it not affect those basic words mentioned above? Could it be that words such as fácil, forma, etc are late loanwwords from Latin? (I would be surprised to learn that, since they are so basic words)

10

Some of these words were re-loaned from Classical Latin after the change of Old Spanish /f/ to /h/ had stopped: compare loaned forma "shape" against inherited horma "mold" (as you mentioned in the question), loaned fácil "easy" against inherited hacer "do", loaned fatal "fatal" against inherited hado "fate".

It might seem surprising that such basic words were loaned, but consider: English "form" and "fatal" are loanwords too, and have become common and well-integrated despite inherited "shape" and "deadly".

The sound change also only happened before a vowel, which is why you see fiesta and fuerte with /f/: the following /j/ and /w/ blocked the change. These words were inherited directly, and were just passed over by the change, which is why there's no doublet *hiesta or *huerte. (It was also blocked by a preceding consonant, hence hacer next to satisfacer, but there are fewer examples of this; Latin /f/ was almost always word-initial for historical reasons.)

8
  • It does seem surprising indeed that such basic words were loaned when spanish was descending from latin itself! Interesting. Do you happen to know where to find the dates (approx) of when was this sound change opperating and/or when were these new loanwords introduced?. Looking at some sample texts from medieval spanish (e.g, Cantar del Mío Cid), it looks like it had only half-happened back then (you find ha but also fablar). By the way, both hada and hado exist in modern spanish, I guess descending no doubt from latin Fata and fatum – Qwertuy Jan 1 at 19:06
  • @Qwertuy Oh, interesting; I wasn't able to find it in my dictionary but Wiktionary does indeed list it. I'll change that in my answer. – Draconis Jan 1 at 19:07
  • 2
    @Qwertuy Also, it looks like the loans caught on in the 16th century or so; that's when we start seeing F and H used distinctly instead of interchangeably (because before that there was no distinction at the start of words, so no reason not to spell hacer with an F even if it was pronounced with /h/). It's harder to tell when the sound change was happening, except just "before that"; some words started being spelled without their initial F around the ninth century so that's a reasonable estimate. – Draconis Jan 1 at 19:15
  • 2
    Isn't satisfacer a partially learned form? The satis part at least does not follow the usual Latin sound changes of intervocalic [t] > [d] and short [i] > [e] – brass tacks Jan 2 at 5:10
  • 1
    @brasstacks Maybe; I didn't treat it as a loan mostly because we see changes like the ch in the participle. I need to find some loaned verbs with initial F to see if they show that same change. – Draconis Jan 2 at 5:14
7

The occurrence of the sound change [f] > [h] > ∅ in modern Spanish words does seem fairly unpredictable. I think this is a situation where dialect mixing and reborrowing/learned re-formation of words caused a lot of complications.

Conditions of the sound change

As far as I know, words that had the cluster /fr/ in Latin never exhibit this sound change. So we see fregar < fricāre, frío < frīgidus.

Words with the cluster /fl/ show in some cases a change to ll word-initially, or ch after [n] (the same outcome as other clusters of a voiceless obstruent + [l], such as [kl] [pl]). Examples:

  • llama < flamma
  • hinchar (there's also something odd going on with the initial "h", which Wiktionary says is from metathesis) < inflāre

There are also words with [fl], such as flama and flor. Furthermore, Ralph Penny mentions that in lacio < flaccidus, Latin [fl] shows the outcome [l] in modern Spanish (A History of the Spanish Language, 2.5.7), although this seems to be the least common outcome.

Latin did not have words starting with [fj] or [fw]. These sequences mostly developed in Spanish from diphthongization of mid-low vowels *ɛ and *ɔ. Some words show loss of [f] in this context, such as hierro < ferrum, huelga < holgar < follicāre, huesa "grave" < fossa. There are apparently further examples in Old Spanish of words with [f] in Latin being spelled with hie and hue.

But currently,

  • words from *fɔ mostly have [fwe], such as fuego < focus, fuelle < follem, fuente < fontem, and others.

  • words from *fɛ mostly have [fje], such as fiero < ferus. There is also a [fje] word fiel where [je] comes from Latin [iˈdeː].

Learned influence

One of the words you mention, fatal, shows another exception to a Latin to Spanish sound change: the lenition of intervocalic [t] to "d". So it seems reasonable to suppose that this word is a learned/loaned form.

I believe some words show partial influence of Latin forms mixed in with some Latin to Spanish sound changes, but I don't know enough to say what the history of the words forma, fácil, fuerte is.

Dialect mixing

Something else to keep in mind as a possibility (I'm not sure how frequently it might have been important) is "horizontal" borrowing from another Romance language that did not have [f] > [h] as a regular sound change.


Another note on dialects: I have read that some dialects of Spanish show [h] or a similar sound where Castilian has ∅ or [f]. The pronunciation with [h] may correspond to a spelling variant with the letter "j". For example, jediondo (vs. hediondo) < *foetibundus.


Links to some discussions of this sound change:

2
  • Interesting. Spanish has flácido and fétido, too, by the way. Perhaps also late loanwords reintroduced directly from Latin? Otherwise, they look odd, specially taking into accounts your examples hediondo and lacio. – Qwertuy Jan 3 at 20:48
  • 1
    @Qwertuy Yup, those are almost certainly learned words too; we’d expect something like *hedio if fētidus had been inherited directly. – Janus Bahs Jacquet Jan 4 at 13:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.