For the longest time I thought the earliest preserved published, or written work, in Spanish was "El Cantar del Mío Cid" ("The Poem of the Cid"). However, I ran today into the accepted answer to question 17145 that mentioned "Glosas emilianenses" as the oldest known text in Spanish ("el texto más antiguo conocido en lengua castellana").

Is this a contradiction? What is currently the accepted oldest publication in Spanish? Are we calling the "Glosas emilianenses" simple the oldest written Spanish we know about?

Update: In deference to comments below, I would like to define "Spanish" as the language first known as "Castellano", and I am looking for the first or earliest preserved published work in Spanish.

I am aware there is a copy of "El Cantar del Mio Cid" from the 14th century en La Biblioteca Nacional de España (National Library) in Madrid, Spain.

  • 7
    First you have to define Spanish. That's not super easy to do in the sense that there's no singular accepted definition :) If we're talking just writing, we can find inscriptions in Romance/Neolatin/Castilian/Cantabrian dating back a really, really long way basically to the Visigoths Oct 3, 2016 at 17:49
  • 2
    The Glosas are not actually a published work, but annotations at the margin of another book.
    – rodrigo
    Oct 3, 2016 at 17:52
  • @guifa, I will use the definition as explained in the answer to question 913. Since we are referring to the earliest publication, Castellano and Spanish, would be interchangeable, I believe.
    – David
    Oct 3, 2016 at 18:10
  • @rodrigo, correct. I am interested in the first preserved published work.
    – David
    Oct 3, 2016 at 18:11
  • 3
    @David I don't think you get what I was meaning. Consider the case of English. At what point do we declare a text English as opposed to a predecessor? Do we consider Anglo-Saxon to be English, even though it (a) also gave birth to Scots and (b) is completely unreadable for modern speakers? The same issue occurs with Spanish, but given that Romance languages exist in a dialect continuum, it's hard to declare a text to be Spanish/Castilian as opposed to one of the other Iberian languages like Galaico-Portuguese, Leonese, Aragonese…, spoken at the time, as they were all much closer then than now. Oct 4, 2016 at 2:31

2 Answers 2


Languages evolve slowly, and Spanish evolved (mainly) from Latin. There are many shades of gray, and it's not easy to determine without ambiguity if some old text corresponds to what we call Spanish. Your definition

I would like to define "Spanish" as the language first known as "Castellano"

is not very useful in this regard. Actually, the language spoken in Spain between (say) 1000 and 1500 is known as Old Spanish

Some (respected, it seems) scholars argue that "Noditia de kesos", an inventory of cheese from year 980, is the most antique text that reflects in written form the new spoken languages that were replacing Latin in the common people speak. This can hardly be called Old Spanish, or even less "Castillian", though (" it represents an Iberian Romance predating any useful distinction between Leonese and Castilian" ref).

Other argue that the Glosas Emilianenses conserves the oldest text in Old Spanish, but this is also disputed ("There is still some debate as to whether the Iberian Romance language of the glosses should be classed as an early form of Castilian or of Aragonese, although some recent studies show that most features belong indeed to the latter.")

In any case, the Cantar of Mio Cid (around 1150) is, if not the oldest, the oldest that has relevance, both for its length, quality and influence.

  • thank you for your answer. I find it very interesting. [+1 for the "Noditia de kesos" reference]. In my defintion of Spanish/Castellano I was trying to follow RAE's concept that the name castellano is also used when referring to the common language of the state in relation to other co-official languages in their respective autonomous territories, such as Catalan, Galician, and Basque. See Is there a difference between “español” and “castellano”?
    – David
    Oct 4, 2016 at 19:32

In the book Manuale di linguistica e filologia romanza (Il Mulino, Bologna, 3rd edition from 2009), by Italian linguists Lorenzo Renzi and Alvise Andreose, there is a section devoted to the first texts in Romance languages. The subsection dedicated to Castilian shows how a variety of Spanish texts were written before Cantar de mio Cid, not only Glosas emilianenses. The importance of Cantar de mio Cid is that is the first literary work of non religious argument that was completely written in the Castilian language. I found especially interesting the last part of the following quotation, which talks about texts written in Mozarabic Spanish.

This is my translation (original text at the end of the post):

The oldest document in Castilian is a text of a practical nature, the Nodicia de kesos, a list of cheeses (precisely kesos, today written quesos) consumed in the monastery of saints Justo y Pastor (in León) which was transcribed on the recto of a document of 959 and slightly later. However, this text remains substantially isolated in the Iberian panorama. From about a century later (about 1090) is another list of goods from Aragon, the Particigon que feci senigor Sango Garcece ('Subdivision that made Sancho Garçes'), which perhaps was intended for the preparation of a legal text. In the documentary field, the vernacular began to emerge in Latin acts during the XI century, and became more frequent in the second half of the century, as shown by some documents from the monastery of San Juan de la Peña, datable to 1062-1063. But it is still an involuntary, unintentional use, to be ascribed to the inexperience and lack of familiarity of the writers with Latin rather than the desire to faithfully record sentences in Romance. The first entirely vernacular documents appear at the end of the XII century, and began to be more frequent during the following century.

Vernacular writings of a religious nature are more frequent, but not much. [...] These are additions of varying length to Latin religious texts made in order to help their understanding: they therefore document the poor knowledge of Latin in the northern regions of Spain between the X and XI century, and not, as in other Romance areas, the use of the vernacular in religious activities. We recall the Glosas Silenses ('Silense glosses', from the Benedictine monastery of Santo Domingo de Silos, in northern Castile) and the Glosas Emilianenses ('Emilianense glosses', from the monastery of San Millán de la Cogolla, in Navarre): the former glosses a Latin Penitential at various points and can be dated to the second half of the X century; the latter translates terms or short passages of patristic texts of various kinds (exempla, sermons) and probably dates back to the XI century.

It was not until the second half of the XII century that two religious compositions appeared in verse, fully comparable, this time, to those already seen in northern France, Provence and Italy: the Auto de los Reyes Magos and the Debate del alma y el cuerpo. The Auto (or Representación) de los Reyes Magos (mid or late XII century) is a liturgical drama consisting of 147 verses of diverse metres with rhyming couplets that tells the story of the three Magi. [...]


The Debate del alma y el cuerpo ('Dispute of soul and body') is a didactic composition, coming from the clerical environment. It remains in fragmentary form (74 seven-syllable verses with rhyming couplets) and can be dated to the end of the XII century. The text – which stages the dispute between the soul and body of a man after his death – derives from French models, and fits into the dialogical genre of disputations (lat. disputationes, fr. débats) that in Latin and Romance Middle Ages were widely spread.

Except for some short texts that have come to us indirectly from later works, the first vernacular literary text of profane subject is the Cantar de mio Cid, the oldest and most important Spanish epic poem, which, although transmitted to us from a XIV century manuscript, was composed – according to the latest theories – in 1207 by a certain «Per Abbat», who probably reworked and assembled pre-existing materials. [...]


Ancient, and very interesting, are some short texts in Romance that come from the central-southern part of Spain dominated by the Arabs (al-Andalus), where there was also a strong and important Jewish community. From the mid XI century, some verses written in a language different from the rest of the composition begin to be found in the final part of the last stanza of numerous muwaššhāt (sing. muwaššhā, literally 'decorated with a shawl'), that is, poems in Arabic or classical Hebrew, which are called harǧāt (sing. harǧā, 'final exit'). Harǧāt are generally written in vernacular Arabic, but there is a large number written in Romance. These are samples of the Mozarabic variety of Spanish [...], which are transcribed – as the rest of the poems – in Arabic or Hebrew characters. This circumstance made that these texts remained unknown to the Romanists until when, in 1948, an English scholar of Semitic literatures, Samuel Stern, deciphered them. Harǧāt – generally consisting of short monologues of a loving subject pronounced by a girl of the people – in all probability derive from poems preexisting to muwaššhāt, and prove the existence of a traditional Iberian poetry of popular origin.

These testimonies of southern Spanish –later extinct in favor of varieties imported from the north of the peninsula– are exceptional in the Romance landscape. The situation of Arab Spain was very different from that of France and Italy: it was an economically thriving, culturally refined and dynamic area. The fact that the literate class (except for Christian religious) did not use Latin as a cultural language prevented that situation of diglossia written Latin / spoken Romance which in other Romance areas had long prevented the writing of the vernacular: in the context linguistic and cultural aspects of Arab Spain, on the other hand, the perception of specificity of the language of the people (Mozarabic) compared to the high language (Arabic and Hebrew) was facilitated by the radical differences between the two codes. The use of Spanish in poetry was linked to the exceptional circumstance of a new literary genre which had foreseen the presence of one female voice of the people. Thereafter, the development of poetry in Spanish will be blocked – at least until the end of the XV century – by the adoption in Castile of the Galician-Portuguese as the specific language of love poems [...].

Note that, after the above cited book was written, the Spanish Royal Academy has declared the Cartularies of Valpuesta, whose oldest documents date back to the IX century, to be the oldest texts featuring the earliest words written in Spanish.

The original of the quoted text is this:

Il più antico documento del castigliano è un testo di carattere pratico, la Nodicia de kesos, un elenco dei formaggi (kesos appunto, oggi scritto quesos) consumati nel monastero dei santi Justo y Pastor (nel León) che fu trascritto sul recto di un documento del 959 e di poco posteriore. Tale testo rimane però sostanzialmente isolato nel panorama iberico. Di circa un secolo posteriore (1090 ca.) è un'altra lista di beni proveniente dall'Aragona, la Particigon que feci senigor Sango Garcece ('Suddivisione che fece il signore Sancho Garçes'), che forse era destinata alla preparazione di un testo giuridico. In ambito documentario, il volgare comincia ad affiorare in atti latini nel corso del secolo XI, e si fa più frequente nella seconda metà del secolo, come mostrano alcuni documenti provenienti dal monastero di San Juan de la Peña, databili al 1062-1063. Ma si tratta pur sempre di un uso involontario, non intenzionale, da ascriversi all'imperizia e alla scarsa dimestichezza degli scriventi col latino piuttosto che alla volontà di registrare fedelmente delle frasi in romanzo. I primi documenti interamente in volgare compaiono alla fine del secolo XII, e s'infittiscono nel corso del secolo successivo.
Più frequenti, ma non di molto, le scritture volgari di carattere religioso. [...] Si tratta di aggiunte di varia lunghezza a testi religiosi latini fatte allo scopo di aiutarne la comprensione: documentano quindi la scarsa conoscenza del latino nelle regioni settentrionali della Spagna tra X e XI secolo, e non, come nelle altre aree romanze, l'uso del volgare in attività religiose. Ricordiamo le Glosas Silenses ('Glosse silensi', provenienti dal monastero benedettino di Santo Domingo de Silos, nel nord della Castiglia) e le Glosas Emilianenses ('Glosse emilianensi', dal monastero di San Millán de la Cogolla, in Navarra): le prime glossano in vari punti un Penitenziale latino e sono databili alla seconda metà del X secolo; le seconde traducono termini o brevi passi di testi patristici di varia specie (exempla, sermoni) e risalgono probabilmente all'Xl secolo.
Solo nella seconda metà del XII secolo compaiono due componimenti religiosi in versi, del tutto paragonabili, questa volta, a quelli già visti in Francia del nord, Provenza e Italia: l'Auto de los Reyes Magos e il Debate del alma y el cuerpo. L'Auto (o Representación) de los Reyes Magos (metà o fine del XII sec.) è un dramma liturgico costituito da 147 versi di varia misura a rime baciate che narra la storia dei tre Magi. [...]
Il Debate del alma y el cuerpo ('Disputa dell'anima e del corpo') è un componimento di carattere didascalico, proveniente da ambiente clericale. Si conserva in forma frammentaria (74 settenari a rime baciate), ed è databile alla fine del XII secolo. Il testo – che mette in scena la disputa tra l'anima e il corpo di un individuo dopo la sua morte – deriva da modelli francesi, e si inserisce nel genere dialogico dei contrasti (lat. disputationes, fr. débats) che nel Medioevo latino e romanzo conobbe un'ampia diffusione.
Se si eccettuano alcuni brevi testi tramandati indirettamente da opere più tarde, il primo testo letterario volgare di argomento profano è il Cantar de mio Cid, il più antico e importante poema epico spagnolo, che, per quanto tramandato da un manoscritto trecentesco, fu composto – secondo le ultime teorie – nel 1207 da un tal «Per Abbat», che probabilmente rielaborò e assemblò materiali preesistenti. [...]
Antichi, e molto interessanti, sono alcuni brevi testi in romanzo che provengono dalla parte centromeridionale della Spagna dominata dagli Arabi (al-Andalus), in cui era presente anche una forte e importante comunità ebraica. A partire dalla metà dell'XI secolo, si incominciano a trovare nella parte finale dell'ultima strofa di numerose muwaššhāt (sing. muwaššhā, lett. 'ornato di scialle')' cioè di poesie in arabo o ebraico classico, alcuni versi scritti in una lingua diversa da quella del resto del componimento, che prendono il nome di harǧāt (sing. harǧā, 'uscita finale'). Le harǧāt sono scritte in genere in arabo volgare, ma ne esiste un cospicuo numero in romanzo. Si tratta di campioni della varietà mozarabica dello spagnolo [...], che sono trascritti – come la parte restante delle liriche – in caratteri arabi o ebraici. Tale circostanza ha fatto sì che questi testi rimanessero ignoti ai romanisti fino a quando, nel 1948, uno studioso di letterature semitiche, l'inglese Samuel Stern, non li ha decifrati. Le harǧāt – costituite in genere da brevi monologhi di soggetto amoroso pronunciati da una fanciulla del popolo – derivano con ogni probabilità da liriche preesistenti alle muwaššhāt, e provano l'esistenza di una lirica tradizionale iberica di origine popolare.
Tali testimonianze dello spagnolo meridionale – poi estinto a favore delle varietà importate dal nord della penisola –sono di carattere eccezionale nel panorama romanzo. La situazione della Spagna araba era molto diversa da quella della Francia e dell'Italia: era un'area economicamente fiorente, culturalmente raffinata e dinamica. Il fatto poi che la classe alfabetizzata (fatta eccezione per i religiosi cristiani) non usasse come lingua di cultura il latino impedì quella situazione di diglossia latino scritto/romanzo parlato che nelle altre aree romanze aveva per lungo tempo impedito la scrittura del volgare: nel contesto linguistico e culturale della Spagna araba, invece, la percezione della specificità della lingua del popolo (il mozarabico) rispetro alla lingua alta (l'arabo e l'ebraico) fu facilitata dalle radicali differenze tra i due codici. L'uso dello spagnolo nella lirica era legato alla circostanza eccezionale di un genere letterario nuovo ehe aveva previsto proprio la presenza di una voce femminile del popolo. In seguito, lo sviluppo eli una lirica in spa­gnola sarà bloccato – almeno sino alla fine del XlV secolo – dall'adozione in Castiglia del galego-portoghese come lingua specifica della poesia di argomento amoroso [...].

  • Please, feel free to improve my English translation.
    – Charo
    Feb 2, 2020 at 22:21
  • 1
    Now, it is much better...ufa. :)
    – Lambie
    Feb 3, 2020 at 19:49
  • 1
    @aparente001: Italian corresponds to the original text: I think it's better not to omit the original. Anyway, you don't need to read it.
    – Charo
    Feb 4, 2020 at 10:37
  • 1
    Gran trabajo Charo! Me permití la libertad de mover la edición en inglés al principio, para que aparezca primero lo que la mayoría de usuarios entenderán. Si no estás de acuerdo, revierte, claro está.
    – fedorqui
    Feb 4, 2020 at 13:17
  • 2
    I think your English version is fine. As a very poor reader of Italian I would not dream of correcting it anyway but it reads well. Thank you for bringing it to our attention.
    – mdewey
    Feb 4, 2020 at 13:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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