Liquid s

We Spaniards have a curiously stubborn inability to correctly pronounce foreign words beginning with /s/+consonant, such as, for instance, "Spain", which invariably we utter as Ehs-pain, to the amusement of English speakers. It is only by training ourselves that we manage to get rid of that spurious "e" preceding the "s" sound, but somehow seems like adding it comes naturally to Spanish speakers. How come we universally apply this phonetic transformation /s/+consonant > /es/+consonant without having been explicitly taught to do so?

In all Iberian Romance languages (and seemingly early stages of other Romance languages like French), Latin words beginning with the so-called liquid s ("s líquida") evolved to include a supporting "e" preceding the "s" sound, vg. scribere > escribir, spatium > espacio, etc. As a result, Spanish (and Portuguese, Catalan, etc.) lexicon does not have any words beginning with /s/+consonant, and this phonetic construct just does not belong in Spanish phonotactics. Consequently, when confronted to it, the Spanish speaker will try a transformation to a valid sequence within the Spanish phonetic system. The question is why the transformation /s/ > /es/ is invariably chosen instead of other valid candidates like /s/ > /as/, /s/ > /is/, etc. I don't have a definite answer, but the following hypotheses come to mind:

Hyp1. Somehow we have been taught to read "s"+consonant (at the beginning of a word) as /es/+consonant, in the same way as we learnt how to read out other letter combinations, regardless of whether they correspond to actual words or not.

Hyp2. This is just a manifestation of the ancient phonetic rule that wiped Latin liquid esses: much as our folk were aware of that transformation we also are and thus use it to deal with alien phonetics.

Hyp3. /es/+consonant is more frequent in Spanish than the formations with the rest of the vowels and diphthongs, so we choose it following a sort of Bayesian argument.

Hyp4. We are influenced by the pronunciation of related Spanish words, which in most cases carry the supporting /e/ as a result of the ancient phonetic transformation from Latin liquid s. So, when first confronted to the word "state", the Spanish words "estado", "estatal", etc. appear as associations in the mind of a Spanish speaker, which induces the application of the /s/ > /es/ rule.

Hyp5. The only plausible candidate transformations are of the form /s/+consonant > vowel+/s/+consonant or /s/+consonant > diphthong+/s/+consonant. Of the five vowels and about a dozen diphthongs in Spanish, /e/ is perceived by speakers as a neutral or default vocalic sound (a role possibly played by /ə/ in English), hence we choose it instead of the others.

Hyp6. /es/ is perceived by the Spanish speaker as the perceptually closest construct to the ideal /s/. That is, [espein] is perceived as closer to [spein] ("Spain") than, say, [aspain].

A short discussion of each hypothesis:

Hyp1. This does not look much reasonable, since Spaniards' reading learning systems certainly never deal with words beginning with the letter "s" plus a consonant, and it is unlikely that most Spanish infants are exposed to foreign words outside their formal reading training and taught to pronounce them by some adult.

Hyp2. Again, this does not seem to hold water. Phonetic rules are not part of a language transmitted knowledge: infants are not actively taught the language, merely exposed to it, so all that they can acquire during the process are the results of such phonetic rules. The illusion of a phonetic rule A→B being transmitted from generation to generation and slowly applying their way to the sounds of a language stems from the dynamics of coexisting subpopulations favoring either A or B. Once A has been completely removed, no remnants of the rule persist on future speakers' minds. The elimination of Latin liquid s having completed centuries ago (possibly even before the formation of Spanish), we cannot invoke such ancient process to explain the current phenomenon.

Hyp3. This is not an entirely illogical argument, but its cogency is hard to estimate. Certainly /es/+consonant is much more frequent in Spanish than the rest of candidates, but not overwhelmingly so.

Hyp4. This hypothesis is very plausible and, in my opinion, must play a role at least in some cases, those where obvious equivalent words exist in Spanish. The problem is that the rule is also applied when there are no equivalent Spanish words to resort to, as I have verified by challenging people without knowledge of foreign languages to pronounce /s/+consonant-beginning words without similar-looking correlates in Spanish: it was also /e/ in those cases that was added to the pronunciation.

Hyp5, Hyp6. These explanations are very similar but show some differences: Hyp5 presents the case of a speaker trying to pronounce no vocalic sound and instead failing and uttering the default /e/, while in Hyp6 the speaker is trying to mimic a previously heard sound. Hyp6 is backed by examples of comparable situations in other languages, like the notable case of English pronunciation by Japanese speakers, featuring a kind of vowel-inserting transformation called anaptyxis (for instance, "strike" is pronounced something like [sutoraiku]). However, I lean towards Hyp5, because the /s/ > /es/ phenomenon also shows when reading a word rather than hearing it, even among speakers with negligible exposure to foreign languages, as I commented above. In favor of Hyp5 there is also the fact that the spurious /e/ is also included in a more exotic context, those of African words beginning with /n,m/+consonant, like Ntutumu, Nguema, Mbasogo (common personal names in the former Spanish colony Equatorial Guinea), which current Spaniards almost exclusively access to in written form.

I would love to know others' opinions on this issue --particularly from those with expertise in Phonetics.