The term ‘Old Spanish’ is widely used in the syntactic literature and indeed the linguistic literature more generally, as a counterpoint both to ‘modern Spanish’ and also to names for other varieties of Old Romance, such as ‘Old Portuguese’, ‘Old French’ and ‘Old Catalan’. The label is a useful one and it occurs frequently in this book, together with the term ‘medieval Spanish’, with which it can be regarded as being extensionally equivalent. It should nevertheless be noted that for most of the period of its existence, what we now think of as Old Spanish would not actually have been referred to as español (or any ancestral form of this word). Indeed, as late as the thirteenth century, usage does not point to the existence of any agreed name for the language. For example, in the prose manuscripts of the Alfonsine corpus, designations range from plain romanz or romançe through to lenguage castellano, romanz de Castiella and even the modern-looking castellano.¹ To the extent, then, that there was any generalized awareness on the part of speakers that they spoke a particular language, the relevant linguistic identity was conceptualized in terms of a Castilian way of speaking or writing. Referring to the language of the medieval kingdom of Castile as ‘Spanish’ thus involves the retrospective application of a modern nomenclature.

This approach should not necessarily be regarded as being ahistorical, however. For modern diachronic syntactic theory (see e.g. Roberts 2007) envisages language change as being realized in specific areas of the grammar rather than in the language as a whole. Moreover, these localized changes, as externally manifested, are long-term processes, evolving as gradual curves (Kroch 1989) rather than in discrete stages. At any given time, therefore, the different components of a language’s grammar will be at different evolutionary moments, implying, logically, that no particular temporal slice in a language’s overall history has any preferential claim to a specific identity of its own. Thus the use of a single term for the entire, seamless continuum is actually well motivated, and for that particular role, it makes sense to employ the familiar, modern name. Dissociated in this way from any linkage to a specific historical period, the word ‘Spanish’ comes to refer not just to the language of the post-medieval Spanish state, but also to all of its previous incarnations. The secondary label ‘Old Spanish’ should accordingly be seen as a meronym, designating no more than a part or segment of the diachronic whole.

This perspective immediately invites the question of what temporal boundaries should be assumed to define the relevant segment of the continuum. Approaching this issue in the first instance with a comparison, it can be noted that what linguists have in mind when they talk about Old Spanish begins rather later than the language variety that falls under the label ‘Old English’. For while the term ‘Old Spanish’ is not generally used to refer to linguistic data which significantly predate the onset of the High Middle Ages, Old English is identified as the language of the Anglo-Saxons, who settled in Britain from the mid-fifth century. Indeed, despite the formal similarity of their names, Old Spanish and Old English occupy different positions within their respective genealogies. In Spanish terms, the nearest counterpart to Old English would be something like late spoken Iberian Latin, i.e. what Wright (1994) refers to as early Ibero-Romance. In practice, Middle English is a better match for Old Spanish, even if, as is implied by the discussion below, the notional beginning of the latter predates that of the former by about two centuries.²

As with other Romance varieties, the Spanish case is complicated by the relatively late appearance of a customized way of writing the language. In its recognizable written guise, Old Spanish effectively dates from the thirteenth century, texts prior to this period tending to be Latinate in appearance.³ At the spoken level, how far back in time what we might be prepared to think of as Old Spanish goes is anyone’s guess. Spanish philological tradition takes its cue from Menéndez Pidal’s seminal work Orígenes del español (1926), a detailed reconstruction of the early spoken language based on documents from the tenth and eleventh centuries.⁴ Linguistically speaking, there is no particular reason to identify those centuries as marking the emergence of a new linguistic entity. However, there is a fairly widespread assumption that they represent, in some way, the época de orígenes ‘origins period’ of the Spanish language. The more fundamental point is that there is a disjuncture between Old Spanish as manifested with full clarity by the bespoke writing system that came on stream in the thirteenth century and Old Spanish as an older but largely presumptive linguistic variety, revealed to us through the prism of a written code devised originally for the speech of many centuries earlier.

While the term ‘Old Spanish’ in principle covers both of the linguistic manifestations just highlighted, the second one can be referred to more specifically as pre-literary Spanish, where ‘literary’ alludes to writing in general rather than to literature specifically. This latter form of Spanish is itself presented to us in a wide variety of guises, close approximations to Latin grammar and spelling characterizing one polar extreme and innovative experimental forms the other. In between we find texts which, to varying degrees, mix Latin words and case endings with syntactic structures, vocabulary and spelling patterns that clearly belong to Old Spanish. Menéndez Pidal’s Orígenes del español references both the experimental glosses associated with the San Millán and Santo Domingo de Silos monasteries and a variety of documents embodying the hybrid text type. The latter, it has to be said, point to the existence of a spoken language whose syntax in most major areas is very similar to what we find in the linguistically more transparent manuscripts of the thirteenth century. This is hardly surprising, given that in the majority of cases Menéndez Pidal’s documents pre-date those of the early literary period by no more than two hundred years, a timeframe which realistically does not allow for any dramatic transformation of the grammar.

The similarity in syntax between the presumptive spoken language of the época de orígenes and the well-attested Spanish of the thirteenth century is illustrated in the extract in (1) below, which is taken from a document originating in the Palencia area and written, according to Menéndez Pidal, in 1097.⁵

1. (1)
   Et si ego mici mortem ante uobis uenerit, si de mea ereditate comodo et demeo ganato, aueatis uos jlas duas partes, et jla tertja, siue de ereditate comodo et deganato, jntre promea anima asancti Zoili. Et siuobis uiro meo aut germanis meis jla tertja quesieritis recolere, aprecient jla quantum ualere, et date precio pro jlas duas partes, et jla tertia lexola por amor de Dios; (Archivo Histórico Nacional de Madrid, San Zoil de Carrión P–7; Menéndez Pidal 1926: 35)

   ‘And if death comes to me before you, of my land as with my cattle, you shall have two parts, and the third, be it land or cattle, goes for my soul to San Zoil. And if you my husband or my brothers want to recover the third, its value should be established, then pay a price for two thirds of it, but the remainder I give for free.’

If one looks past the Latin interference, such as the anachronistic dative forms mici (Classical Latin: mihi) and uobis or outmoded spellings like aut, pro and comodo (Classical Latin: quomodo), the language in the extract should seem very familiar to anyone versed in the grammar of the post-1200 period. Rather obviously, the consistent use of jla and jlas to introduce noun phrases points to a fully operational definite a...