In more formal or published texts—including this post—readers expect polished prose free of misspellings, abstruse formatting, awkward phrasings, errant punctuation, or grammatical infelicities. At best, such blemishes distract from content; at worst, they reflect poorly on the writer, editor, and publishing venue, potentially with palpable social consequences (e.g., Collins and Blot 2003:67–120; Jaffe and Walton 2000:561). This ideology of textual presentation is a key motivation behind the philological tradition of publishing edited transcriptions of older manuscripts. Modernization and standardization of orthography, spelling, and grammar, along with transliteration into type, make these works more accessible to a broader audience without the linguistic and paleographic training required to interpret more opaque writings. Content becomes easier to interpret once the hurdle of deciphering an unfamiliar graphic form has been cleared.
But by eliminating internal variation and irregularity, this same practice obscures critical data that could provide insight into phenomena such as the context of composition or the author’s identity. The Título de Quetzaltenango y Momostenango, a Spanish-language título ('title') published in Land, Politics, and Memory in Five Nija’ib’ K’iche’ Títulos, features grammatical errors whose traces were obscured in later transcriptions of the text. Yet closer examination reveals a patterning that indicates the document was composed by a speaker of K’iche’ Mayan with an imperfect command of Spanish. Paleographic and philological scholarship often addresses inconsistencies in spellings as evidence of phonological developments or poor orthography (Lapesa 1981:370–391; Muñoz y Rivero 1889:105–107). However, as the following brief analysis illustrates, the aberrant spelling patterns in the Título de Quetzaltenango y Momostenango reflect grammatical errors revealing linguistic interference through Spanish-K’iche’ language contact (compare for Nahuatl, e.g., Karttunen 1998:428; Karttunen and Lockhart 1976).
Grammatical gender is a lexical property of all Spanish nouns, often overtly marked by an –o ending on masculine forms or –a on feminine forms. Exceptions like the feminine noun la mano (‘hand’) must be acquired with the individual lexeme and tend to be learned later than nouns whose gender is overtly marked (Hernández Piña 1984:235–237). Gender agreement must be marked on all modifiers in a noun phrase, including determiners and adjectives, according to requirements of the grammar (see RAE 2009:81–126; Torrens Álvarez 2007:84–88). In at least twelve cases in the Título de Quetzaltenango y Momostenango, however, nouns and their associated adjectives or articles demonstrate grammatical gender disagreement (compare Bricker 2000:106). This trend results in constructions such as la *mucho [mucha] sangre (‘the great amount of blood’), in which an adjective is marked as masculine but modifies a feminine noun. Interestingly, such over-generalization of the masculine form when a feminine form would be expected is the most frequent error in constructions with gender disagreement (58%, n = 7 of 12), corresponding to a tendency observed among modern-day Spanish learners (Hernández Piña 1984:235–237; Montrul 2004:54, 80–83; White et al. 2004:128).
The Título de Quetzaltenango y Momostenango also produced errors in grammatical number. In Spanish, the plural is usually distinguished overtly from the singular on nouns and their modifiers with an –s or –es suffix, depending on phonological context. Plurality is indicated in verbal inflection by a set of separate endings whose precise form varies according to person, tense, and mood. As with grammatical gender, number agreement between nouns and their modifiers, and between verbs and their subjects, has been grammatically required throughout the history of Spanish (see RAE 2009:127–130; Torrens Álvarez 2007:82–83).
The Título de Quetzaltenango y Momostenango features 55 nominal constructions with number disagreement, such as *mucho [muchos] indios (‘much [many] Indians’), in addition to 27 verbal phrases. These errors most frequently take the form of plural under-marking, with singular inflection when a plural form would be anticipated (76%, n = 62 of 82). In the other cases, a plural form occurs where a singular interpretation would be expected (24%, n = 20 of 82). Interestingly, plural over-marking occurs with disproportionate frequency in contexts of subject-verb agreement, where it is responsible for 41% of number errors (n = 11 of 27; e.g., *empezaron [empezó] este cacique, ‘they began [he began], this cacique’). In contrast, plural over-marking characterizes just 16% of errors in nominal phrases (n = 9 of 55; e.g., *estas [esta] niña, ‘these [this] girl’).
Frequent errors in gender and number agreement would be surprising in a text produced by a scribe who had learned Spanish natively. However, they would not be unexpected from a native speaker of K’iche’, a language which does not have grammatical gender. Similarly, unlike standard Spanish, K’iche’ does not categorically mark number agreement in nominal or verbal phrases (see López Ixkoy 1997:101). Because overt plural marking is not grammatically required, plurality is often under-marked on K’iche’ nouns, articles, adjectives, or verbs. Errors in grammatical production by a native-speaking K’iche’ writer who had not fully mastered Spanish would be particularly likely in areas in which the K’iche’ and Spanish grammars fundamentally differ, as with lexical gender or number morphology.
In sum, the nature of the gender and number (dis)agreement in the Título de Quetzaltenango y Momostenango suggests that the author was a native K’iche’ speaker who was not sufficiently bilingual in Spanish to suppress the influence of his native language’s grammar. The current evidence is insufficient for evaluating the scribe’s acquisition of Spanish in more detail, including age or learning context. But critical examination of orthographic details emphasizes the importance of preparing reliable transliterations, ideally paired with facsimiles of the original folios, to accompany published transcriptions of earlier manuscripts. This practice not only clearly differentiates the input of the original scribe from that of the scholar, but also more generally allows readers to reconstruct and critically evaluate the philological interpretations underlying these publications. More broadly, this short case study exemplifies the potential of paleographic and grammatical irregularities to contribute to our understanding of historical sociolinguistics. Although secondary to traditional concerns of philological or literary analysis, features of irregular spelling or grammar in original manuscripts can be critical to sociohistorical perspectives, without which our understanding of any text remains incomplete.
Mallory E. Matsumoto is a PhD student in the Department of Anthropology at Brown University. Land, Politics, and Memory in Five Nija’ib’ K’iche’ Títulos is her first book.