The phonetic sound change → (a form of debuccalization), followed by a phonemic restructuring resulting in the complete loss of the sound ( > ), represents a significant development in the phonological history of the Spanish language. This change is also observed in various Romance languages, including Gascon, Aromanian, Moldavian, Transylvanian Romanian and Most Eastern Asturleonese, as well as sporadically in other Romance languages. Under specific phonological conditions, the initial Latin /f/ evolved to [h], which eventually disappeared in standard Spanish. However, its pronunciation persists in some words across certain dialects, particularly in parts of Andalusia, Extremadura, and Latin America. It is also maintained in transitional dialects such as Cantabro and Extremaduran. An example of this phenomenon is the Latin word FARĪNA, which evolved to /aˈrina/ in Spanish (with the retained in the spelling harina) compared to the Italian /faˈrina/ for "flour").

The phoneme /f/ in Latin phonology

The place of /f/ in the consonant system In Latin words inherited from the proto-language, /f/ could only appear in the initial position. In intermediate positions, it was present primarily in borrowings from other languages (e.g., RUFUS, which was likely borrowed from a neighboring Italic language), or in derived words formed by adding a prefix to a word beginning with F (e.g.: DE-FENDERE, CON-FUNDERE). Following the disappearance of /h/, /f/ was Latin's only fricative apart from /s/, leading to its unstable integration within the consonantal system. As a result, it was particularly susceptible to phonetic evolution.

Phonetic variability of /f/ The sound denoted by the grapheme F may not have been strictly labiodental but could have been bilabial . Even if /f/ functioned as an isolated phoneme, it may have had two allophones in pronunciation. Some researchers argue that this characteristic developed in Spanish under the influence of Indo-European languages spoken in the region where the language originated; however, this assertion is difficult to verify. It is likely that in Ibero-Romanic dialects, the bilabial pronunciation was more commonly used.

The labiodental realization of /f/, which appears in languages such as French, Italian, Portuguese, and Romanian, may have emerged through analogy with the change [β] > [v]. In these languages, the originally Latin semivowel /w/ transitioned to [β] and ultimately consolidated into a labiodental /v/. This particular evolutionary phase did not occur in the northern regions of the Iberian Peninsula. If /f/ is assumed to have been pronounced labiodentally, it would not have had a corresponding sound partner, leading to a lack of conformity within the consonant system.

Evolution of /f/ in Spanish

Possible allophones and their distribution The phoneme /f/ can be realized as a bilabial fricative [ɸ], which exhibits variability in its phonetic realization based on surrounding sounds. This articulation may manifest as either fortis or lenis, resulting in three proposed allophones:

before the back vowels /o, u/,

(which can also be transcribed as ) before the semivowel /w/, and

in the other positions; that is, before the vowels /i, e, a/ and the consonants /j, l, r/.

The distribution of allophones can be influenced by phonological conditions, which may reinforce or relax articulation. For instance, /f/ can be affected by neighboring sounds, leading to variations like aspiration. In Gascon, the [h] articulation is generalized across all positions, whereas in Spanish, it primarily occurs before vowels, with some exceptions (notably before the diphthong 'ue'):

FATU > hado (also in Spanish)

FESTA > hèsta, Spanish fiesta

FILU > híu, Spanish hilo or filo

FLORE > hlor, Spanish flor

FRATRE > hray (brother), Spanish fraile / fray

FRUCTU > heruto, Spanish fruto

CONFINE > couhí, Spanish confín

PROFUNDU > prouhoun, Spanish profundo

Similar phonetic changes have been observed in various regions of Latin Europe:

Romanian Rural Dialects: Common occurrences include FERRU > hier (Castilian hierro), FILIU > hiu (Castilian hijo); Calabria, Southern Italy: The sound change manifests as FABA > hava (Castilian haba) and FEMINA > hímmina (Castilian hembra); Brescia, Lombardy, Northern Italy: Examples include FAMEN > ham (Castilian hambre) and FEBRUARIU > hebrer (Castilian febrero); Isolated areas of Sardinia: In these regions, the F- has completely disappeared, leading to transformations such as FOCU > oku (Castilian fuego) and FUMU > ummu (Castilian humo). Most Central and Eastern Asturleonese dialects experience this aspiration sporadically always in front of diphthongs /we/ or /wi/: FOCU > ḥueu, ḥuebu (Castilian: fuego), FOLIA > ḥueya (Castilian hoja), FAGINA > ḥuina. In most central dialects, this is sporadic and a speaker may say both versions indistinctively, thus it is usually not reflected in writing. Eastern Asturleonese dialects of Llanes and Cabrales, examples: Examples include FAMEN > ḥambre (Castilian hambre) and FEBRUARIU > ḥebreru (Castilian febrero).

First written testimonies of the change in the historical Castile The earliest evidence for the phonetic change from /f/ to /h/ or the complete loss of /f/ in historical Castile (including La Rioja), dates back to the 9th century. Notably, in a document from 863, the Latin name FORTICIUS is recorded as Ortiço, and in another from 927, it appears as Hortiço. From the 11th century onward, the instances of this change become more frequent, not only in Castile but also in other regions. The sporadic written documentation suggests that this phonetic change may have occurred earlier in spoken language.

It remains uncertain whether this phonetic realization was widespread across all social classes in Castile. The change may have been predominantly observed among lower social strata, while the educated and more conservative classes may have continued to pronounce [f] or [ɸ] in all positions. Alternatively, the aspiration [h] might have been articulated primarily before back vowels. However, definitive conclusions about the extent of this phonetic evolution cannot be drawn until it is consistently reflected in written records, especially since aspiration was often represented by the grapheme "f-" for many centuries. This phenomenon is exemplified in works such as the Cantar de mio Cid, where the Arabic-deriver preposition hasta (from Arabic ) appears as fasta. Similarly, the Arabic term was adopted into Spanish as alfombra. This suggests that speakers may not have recognized a significant acoustic difference between the pronunciations [f] and [h]. Alarcos Llorach (1951) posits that:Alarcos Llorach, Emilio (1951). "Alternancia de f y h en los arabismos". Archivum 1. 29-41.

Phonological context In Spanish, the aspirated articulation of /f/ as [h] is generalized in all pre-voiced positions. This can be observed in the following examples:

FACERE > hacer FÉMINA > hembra FERRU > hierro FILIU > hijo FOLIA > hoja FUMU > humo Some exceptions exist, often relating to cultisms, including febrero, fiebre, fiesta, filo, fin.

Prefixed words have also undergone similar phonetic changes once they were recognized as such:

OFFOCARE > ahogar SUFFUMARE > sahumar

In other contexts, intervocalic -F- typically evolved into a [β], represented by v or b in writing, following the analogy of the evolution of original voiceless stops:

PROFECTU > provecho RAPHANU (< Gr. ῥάφανος) > rábano

However, instances of complete loss also occur.

DEFENSAM > dehesa

In Medieval Spanish, the sequence -NF- gave -f- (or -ff-), as seen in INFANTE > ifante or iffante, which later standardized as infante in contemporary Spanish.

The F- sound has been preserved before consonants, as well as before the semiconsonant [w]. In certain regions, particularly Andalusia and some areas of Spanish America, it may be articulated as an aspirate or a velar fricative in this position. This can be attributed to the labial nature of [w], which can accommodate the preservation of the labial articulations or :

FOCUS > fuego FONTIS > fuente FORTE > fuerte FUIMUS > fuimos FLORE > flor FRIGIDU > frío

There are rare instances where the FL- group loses the initial F-, such as FLACCIDU > lacio. This suggests that, on occasion, the aspiration [hl-] may have appeared. However, in most contexts, the initial f is retained, indicating that other phonetic factors likely influenced these outcomes, similar to the palatalization that affects the initial CL- and PL- groups.

The change /f/ → [h] and Basque–Latin bilingualism

Arguments in favor of the Basque substratum One widely accepted explanation for the phonetic change from /f/ to [h] in certain Romance languages, particularly Castilian and Gascon, is attributed to the influence of the Basque substratum. This theory, articulated by Ramón Menéndez Pidal, suggests that the Basques, along with the Cantabri and possibly the Iberians—whose languages did not include the sound [f]—substituted it with [h], which they perceived as acoustically similar. This hypothesis is supported by several key points. The earliest documented occurrences of the /f/ to [h] change were found in northern Castile, an area adjacent to regions where Basque was spoken. This geographic proximity implies a potential linguistic influence from Basque speakers on the emerging Spanish dialects. Additionally, the phenomenon also appears in Gascon, which lies on the other side of the Pyrenees. This region, historically inhabited by peoples who spoke Basque or similar languages, further strengthens the argument for a Basque substratum affecting local languages.

While similar changes have been documented in other parts of Latin Europe, the widespread and systematic adoption of the /f/ to [h] change is notably prominent in Castilian and Gascon. This suggests that these areas were more significantly influenced by the ancient Basque–Aquitanian populations, which predated the Roman conquest. The absence of the /f/ sound in the Basque language may have led to a natural phonetic shift in neighboring languages, where speakers adapted their pronunciation in contact situations, favoring aspirated forms over non-existent phonemes.

Objections against the substrate theories While the theory proposing a Basque substratum as a catalyst for the phonetic change from /f/ to [h] has gained some acceptance, it has also faced several objections. One primary concern is the uncertainty surrounding the existence of the aspirated sound [h] in medieval Basque. This raises questions about whether [h] could realistically have replaced [f]. Menéndez Pidal argued that Latin /f/ was labiodental rather than bilabial. However, it is unclear whether Basque speakers were unable to articulate the labiodental [f], particularly given that in certain Basque dialects, what was previously a bilabial fricative evolved into a labiodental [f] in intervocalic positions. Basque philologist Koldo Mitxelena noted that Basques historically did not appear to struggle with producing this sound.Michelena, Luis (1957). "Las antiguas consonantes vascas". Catalán 1957b, 113-58.

Another argument against the influence of a Basque substratum is the preservation of initial /f/ in the Romance dialects of Navarre, an area with a significant Basque-speaking population. If the presence of Basque speakers was a critical factor in the phonetic shift, one might expect to see similar changes in the Navarrese dialect.