Lidiar con fieras: el árabe y el hebreo en el Tesoro de Covarrubias ; Fighting against Beasts: Arabic and Hebrew in Covarrubias’s Tesoro

* The research on which this essay is based was supported by my fellowship at the Mandel Scholion Center at the Hebrew University of Jerusalem, and its completion facilitated by an Arnold Fellowship at the Zvi Yavetz School for Historical Studies at Tel Aviv University. A draft of this essay was presented at a seminar of the CORPI research group (CSIC) in October 2018, where I had the privilege of discussing it with Dominique Neyrod. I am grateful to Mercedes García-Arenal and the members of the group for their valuable feedback, the anonymous reviewers of Al-Qanṭara for their helpful comments and references, and Fernando Rodríguez Mediano, whose generous reading, encouragement and support made the task of writing in the midst of a pandemic possible. AL-QANTARA XLI 2, julio-diciembre 2020 pp. 477-516 ISSN 0211-3589 https://doi.org/10.3989/alqantara.2020.013

The present study examines the representation of Arabic and Hebrew in Covarrubias's Tesoro de la lengua castellana o española (Madrid, 1611) from a comparative perspective. Analyzing the lexicographer's ideological and meta-linguistic discourse, I reconstruct his postulates regarding the respective histories of Arabic and Hebrew, their interrelationship, and place in the history of Spanish. In light of these postulates, and taking into account the lexicographer's access to knowledge, sources, and informants, I examine some of his etymological practices, focusing on his attempts to use grammatical and lexical knowledge of Hebrew to illuminate Arabic etymologies of Spanish words. Combining the ideological and technical levels of the text, I reflect on the interplay of knowledge and ignorance in Covarrubias's treatment of Semitic languages.
to a pure, Castilian vocabulary, avoiding "vocablos agenos y estraños", brought into the language by the "diversidad de gentes estrangeras que en diversos tiempos han venido a nuestra Castilla". 2 Sebastián de Covarrubias y Orozco (1539Orozco ( -1613, who dedicated his Tesoro de la lengua castellana o española (1611) to King Phillip III, declared, in a gesture similar to that of his antecedents, that the purpose of his work was to honor Spain by making readers recognize that its language, albeit vernacular and modern, "no se debe contar entre las bárbaras", but rather equal in dignity to Latin and Greek. As for the Hebrew, Covarrubias went even further than Nebrija or Villalón, and promised to show Spanish was "muy parecida a la hebrea en sus frasis y modos de hablar", creating thereby an affinity that went beyond questions of prestige. 3 The lexicographer's professional trajectory and his written work reflect a deep commitment to the Crown, Church, and the values of Old Christian society. Archival evidence suggests that Covarrubias's father, the Toledo-born poet Sebastián de Horozco, was of converso background or, at the very least, suffered from such a reputation. 4 This dubious lineage, which, technically speaking, made Covarrubias himself a New Christian or suspected of being one, did not stand in his way of reaching high-rank ecclesiastical positions such as the Canonicate of Cuenca; becoming King Philip II's chaplain; serving as a consultor to the Holy Office; or being involved, by mandate of Philip III, in the establishment of schools in Valencia in order to instruct the morisco population in the Kingdom's faith and language between 1596 and 1601. 5 2 Villalón, Gramática, title page and "Prohemio al lector", n.p. 3 Covarrubias, Tesoro, Dedication to the King, n.p. Although I have consulted Arellano and Zafra's edition of the Tesoro and the Suplemento, I quote directly from the 1611 princeps. I modernize punctuation, capitalization, and accentuation, and resolve abbreviations, but respect Covarrubias's original spelling. The only exceptions to this rule are i/j and u/v, adapted to their modern phonetic value in all cases but those in which the lexicographer transcribes Arabic or Hebrew words. Quotations from the Suplemento follow Dopico and Lezra's edition.
Similarly, and perhaps more important for the purposes of this essay, Covarrubias's dubious ancestry did not stop him from writing against conversos and moriscos from an Old Christian perspective, recurring to anti-Jewish and anti-Muslim tropes. He employs a harsh "usvs.-them" rhetoric when referring to New Christians, as one can note, for instance, in his entry "Confesso", a term referring to those converted from Judaism and their descendants, "digamos que confesso es lo mesmo que judío"; or from his entry "Morisco", expressing doubt, if not disbelief, regarding the sincerity of converts from Islam: "si ellos son católicos, gran merced les ha hecho Dios, y a nosotros también". 6

The Quest for Etymology
One cannot overstate the Tesoro's importance to the history of Hispanic lexicography. Unlike Nebrija's lexicon and others written in its mold, the Tesoro, the first monolingual dictionary of Spanish, does not merely present foreign equivalents for Spanish words, but defines them, comments on their usage, documents proverbial expressions, and, at times, provides encyclopedic information related to them. 7 Reading, however, the Tesoro's prologue, it is clear that what the lexicographeras well as his contemporaries-deemed the most significant achievement of this magnum opus was the etymologies it offered. 8 Situating his work within a tradition of Iberian etymological studies going back to Isidore of Seville, 9 Covarrubias explains not only the importance of pursuing the origin of words, but why the history and nature of Language-and in particular, those of Spanish-make this endeavor such an arduous one. Alluding to the title of his Tesoro, the lexicographer 6 Covarrubias, Tesoro, s.v., "Moro", my italics. See also the entries "Marrano" and "Judío" in the Tesoro and Suplemento, analyzed in Reyre, "La voz judío", or the entry "Alcorán". 7 See Alvar Ezquerra, "El Tesoro de Covarrubias". 8 See, e.g., Calero Vaquera, "Apud grammaticos", pp. 164-166; García Macho "La lexicografía monolingüe". 9 In the dedication to Phillip III, the lexicographer refers the edition of Isidore's Etymologiae sponsored by the King's father, Phillip II. In addition, Navarro Arroya's laudatory epistle-printed as part of the Tesoro's prolegomena-refers to Covarrubias's work as Las Etimologías, and mentions a series of etymological works-Isidore's included-insinuating the superiority of Covarrubias's. compares his work to a quest in which, in order to lay one's hands on a treasure hidden in a cave, one must struggle with fierce monsters and beasts who stand at its entrance, wishing to scare one away. After entering-the allegory proceeds-one finds an enchanted Moorish woman ("una mora encantada"), who, sitting on her throne, may or may not grant one access to the treasure. 10 While Covarrubias leaves the ethno-religious identity of the treasure's guardian open for interpretation, 11 he states explicitly that the beasts and monsters in his fable stand for the many foreign languages one needs to master in order to provide a full etymological account of Castilian lexemes-Latin, Greek, Hebrew, Arabic, French, and Italian being some of those he enlists. In a gesture of captatio benevolentiae, he apologizes to the readers who-unlike himself-are not well-versed in Latin, Greek, or Hebrew: Latin appears in the Tesoro with no translation, and Greek and Hebrew appear in their respective alphabets. As for the Arabic, which is always represented in transcription, Covarrubias explains that "en la lengua arábiga casi todos somos iguales, fuera de algunos pocos que la saben, y assí hemos de dar crédito a los peritos en ella". 12 In this statement, Covarrubias not only justifies his particular need to rely on informants, but turns the ignorance of Arabic-an ignorance that is implied, rather than called by its name-into a constitutive element of the Spanish-speaking "nosotros" constructed throughout the text. Compensating for this lack of first-hand acquaintance with Arabic, Covarrubias declares to have used Hebrew as a means of validating the information provided by his informants ("Heme valido de la lengua hebrea para confirmar lo que los susodi-10 "…me atrevo a usar deste término por título de mi obra, pero los que andan a buscar tesoros encantados suelen decir fabulosamente que, hallada la entrada de la cueva do sospechan estar, les salen al encuentro diversidad de monstruos fantásticos, a fin de les poner miedo y espanto para hacerlos volver atrás […] pero venciendo con su buen ánimo y con sus conjuros todas estas fantasmas, llegan a la puerta del aposento, donde hallan la mora encantada en su trono, sentada en una real silla y cercada de grandes joyas y mucha riqueza, la qual si tiene por bien de les dexar sacar el tesoro, van con recelo y miedo…". 11 Opposing interpretations of the tale can be found, e.g., in Calderón, who construes the symbolic act of depositing the treasure in a Moor's hands as a gesture of inclusion or an attempt to integrate the Andalusi heritage with the self-narrative of Spanishness ("Covarrubias y el Tesoro"); and Lezra, who identifies in this "alegoría pseudo-épica" an anxiety centered in the impossibility of conversion, and situated Covarrubias's fable within an exclusive concept of Spanishness ("La mora encatada", esp. pp. 477-482).
12 Tesoro, "Al letor", n.p. chos me interpretan de la arábiga"), gesturing towards a comparative method he employs in order to access what is otherwise inaccessible to him. The Tesoro offers over six hundred Arabic etymologies, and a similar, or slightly higher, number of Hebrew ones, many of which in the same entries. 13 Given the unignorable amount of Spanish words of Arabic origin, and, no less important, the fact that contemporary speakers were aware of this origin even if they had no knowledge of Arabic, 14 the inclusion of Arabisms in the Tesoro is hardly remarkable, even if the methods, sources, or treatment of a language the lexicographer declares to not know still leave much room for discussion. Considering, however, the contribution of Hebrew to Spanish vocabulary-a contribution that, excluding biblical names or the case of Judeo-Spanish, is, at best, anecdotal-the mere presence of Hebraisms, let alone in such numbers, requires an explanation.
In recent years, different aspects of the Semitic presence in the Tesoro-linguistic, cultural, and political-have received a considerable amount of scholarly attention. However, previous studies focused on either Hebrew or Arabic and the representation of the two languages has rarely been addressed from a comparative perspective. A salient exception to this rule is Dominique Neyrod's study, "Relire les étymologies hébraïques à la lumière des étymologies arabes dans le Tesoro", which, examining entries in which both Hebrew and Arabic etymologies are offered, points out to the juxtaposition of different-Hebrew and Arabic-grammatical traditions in the Tesoro, as well as to the lexicographer's use of Hebrew grammatical discourse for understanding equivalent phenomena in the Arabic. 13 Ruhstaller's counts 628 Arabic etymologies in the Tesoro ("Covarrubias como recopilador", p. 318) and Calderón counts 711 ("Covarrubias y el Tesoro", p. 88). Neither one indicates whether their count reflects the number of Spanish lexemes to which an Arabic etymology is offered, or the number of explanations based on an Arabic etymon. As for the Hebrew, Sajó counts 825 lexical units, distributed between 514 Tesoro and Suplemento entries, almost all of which in the context of etymological explanations ("Las etimologías hebreas (i)", p. 125).
14 An oft-cited example for the intuition early modern Spanish speakers had of the Arabic origin of words they used can be found in Cervantes's 1615 Quijote, where Don Quijote explains to Sancho that all Spanish words starting with al or ending with í "son conocidos por arábigos", giving the uneducated squire a long list of examples (DQ2, ch. 67, p. 1285).
The present essay aims to broaden the comparative scope beyond the grammatical domain and examine Covarrubias's treatment of Hebrew and Arabic in relation to one another. Based on the lexicographer's sources and access to knowledge of the two languages, the respective roles they play in his ideological world, and the place they are assigned in his history of Spanish, I examine some of his etymological practices and in particular his attempts to employ comparative methods for the discussion of Arabisms. Combining the ideological, meta-linguistic, and technical levels of the text, I reflect on the interplay of knowledge and ignorance in Covarrubias's representation of Semitic languages.

Sources and Informants
While Covarrubias's declaration of collective incompetence in Arabic may not apply to all sectors of Spanish society, it is undeniable that in the lexicographer's social and intellectual circles, knowledge, training, and reference books of Arabic were incomparably less accessible than those of Hebrew. 15 As Reyre has shown in her studies, Covarrubias received his training in Hebrew as a student in Salamanca between 1565 and 1571. During these years-dramatic times for Spanish Hebraism, and particularly in Salamanca-Covarrubias attended courses imparted by Martín Celanda. In these classes, as Covarrubias himself attests in the Tesoro, the professor would read out loud passages from the Old Testament in Hebrew, and then translate or paraphrase them in the vernacular. 16 While these text-based classes equipped the lexicographer with basic notions of Hebrew morphology, grammar, and phonetics, we have no reason to assume his competence in the language was beyond rudimentary. It did, however, provide him with sufficient tools to find his way around Hebrew grammars and lexicons available in Latin-namely, the Grammatica Hebraea Absolutissima (Antwerp, 1556) by the German convert (and ex-Rabbi) Johann Isaac Levita (to whom Covarrubias refers as Juan Isach), published at least five times during the sixteenth century; the Globus linguae sanctae by the Portuguese theologian Luís de São Francisco (Rome, 1586); and, most important, Sante Pagnini's Thesaurus Hebraicae Linguae, based on David Qimḥi's twelfth-century biblical concordance known as Sefer ha-Šorašîm, and published in various editions throughout the sixteenth century, some of which included a summary of Hebrew grammar. 17 As for Arabic, although we have no reasons to assume that Covarrubias ever studied it, it is worth noting that Arabic, too, occupied some part of the curriculum during his years in Salamanca. As López-Baralt and Martínez de Castilla Muñoz have shown, the Hebraist Martín Martínez Cantalapiedra, who occupied the University's Trilingual Chair from 1543 until his death in 1579-with a five-year hiatus during which he was imprisoned by the Inquisition-dedicated part of his classes to Arabic, using Ibn Ājurrūm's grammatical treatise known as al-ʾĀjurrūmiyya. While testimonies of students attending these classes indicate that this academic framework was insufficient in terms of fostering a reading competence, they also suggest that acquaintance with the Arabic alphabet or basic grammatical principles was not completely out of reach. 18 Be that as it may, the sources used by Covarrubias for the discussion of Arabisms differed from the Hebrew grammars and lexicon in two important aspects: first, they did not focus on Arabic as an object of inquiry in its own right, but rather on its lexical traces in other languages; 19 and second, they represented Arabic words in Latin transcrip-17 Sante Pagnini's Thesaurus, first published in Lyon in 1529, was revised and reedited several times during the sixteenth and seventeenth centuries. According to Reyre, Covarrubias was most likely to have consulted an abbreviated version revised by Arias Montano, titled Thesauri Hebraicae Linguae Epitome (Antwerp, 1572), which also included a grammatical summary ("Cuando Covarrubias", p. 11 note 31). In this article, both the princeps and the abbreviated Antwerp edition were examined. For an overview of Covarrubias's Hebrew sources, see, besides Reyre's aforementioned study, Lépinette, "Contribution à l'étude du Tesoro"; Sajó, "Las etimologías hebreas (i)", pp. 127-128. On the Thesaurus's sources, editions, structure, and grammatical terminology, see Attia, "Aux origines du Thesaurus".
18 López-Baralt, A zaga de tu huella; Martínez de Castilla Muñoz, "Arabic in Salamanca"; and cf. Bataillon, "L'arabe à Salamanque". 19 A possible exception to this rule is Pedro de Alcalá's Vocabulista, which Covarrubias seems to have used in a few isolated cases. See note 83 below, and Ruhstaller, "Covarrubias como recopilador", p. 326-327. tion. Thus, besides a few compilations mentioned throughout the Tesoro, the most notable of which is Francisco López Tamarid's Compendio de algunos vocablos arábigos introducidos en la lengua castellana, the two peritos Covarrubias proclaims to have consulted are Diego de Urrea, the King's interpreter of Oriental languages, and Fray Diego de Guadix, an Inquisitorial interpreter from Granada and author of a treatise titled Recopilación de algunos nombres arábigos que los árabes pusieron a algunas ciudades y a otras muchas cosas (ca. 1593), unpublished at the time. 20 Urrea, whose unique life trajectory was brought to light by Rodríguez Mediano and García-Arenal, was born in Calabria, but at a very young age was captured by the Ottomans, converted to Islam, and educated in the madrasa in Tlemcen, where, as one of the seventeenthcentury historians writes, he studied "la lengua Arábiga y Turca, que escreviéndolas y leyéndolas las hizo más suyas que si fueran naturales: y estudió Gramática Árabe, la Lógica y Philosophía con muchas ventajas, enterándosse muy bien del Alcorán y su maldita Theología, por tanto estremo, que no uvo Moro que le aventajasse." 21 At the age of thirty, Urrea was captured once again, this time by Christians, and after spending a few years in Sicily and re-converting to Catholicism, ended reaching the Iberian Peninsula. His profound, first-hand knowledge of Arabic language and Islamic culture, so difficult to find within the boundaries of late-sixteenth-century Spain, made him not only the royal interpreter of Arabic, Turkish, and Persian, but also the Crown's diplomat and special envoy to North Africa, Arabic professor at Alcalá, and the scholar in charge of the collection of Arabic manuscripts in El Escorial. 22 In contrast, Guadix's mastery of Arabic was not only significantly inferior-as one can learn from his Recopilación-but confined, for the most part, to the dialect spoken around Granada. The friar's attempts at showcasing his skills in fuṣha, manifest in a couple of verses 20 There are two modern editions of Guadix he claims to have copied from one of Ibn Sīnā's books, reveal an overwhelming amount of orthographic errors, incorrectly connected letters, and poor word segmentation. Similarly, his use of Classical grammatical terms reveals an unbearable gap between pretense and actual grasp. 23 Moreover, from an etymological perspective, most of Guadix's propositions cannot be considered plausible even by the criteria of his time, as one can learn, perhaps, from Covarrubias's choice to integrate into his Tesoro less than ten percent of the etymologies suggested the Recopilación. 24 Unlike Urrea, with whom Covarrubias seems to have communicated both in person and in writing, in the case of Guadix, the lexicographer attests to have consulted some of the friar's writings-"he visto algunos escritos"-without specifying their title nor the nature of his reliance on them. While collations of corresponding entries in the Recopilación and the Tesoro reveal a notable resemblance between the texts, minor differences between them suggest, as Bajo Pérez has it, that Covarrubias may have used a different version of the Recopilación. Alternatively, Ruhstaller maintains that Covarrubias had consulted at some point the full text of the Recopilación, but at the time of compiling the Tesoro relied on the notes he had taken rather than the original manuscript. 25 I shall return to this point later.

The Natural Language
Before delving into Covarrubias's ideological discourse on the Semitic languages, let us examine an anecdote in which he expounds his theoretical approach to language in general and reveals his modus operandi as a writer navigating between narrative, ideas, and linguistic material. Early modern polemics regarding the identity of humankind's primitive language and the motivation attributable to national lan-guages often took recourse in the notion of "natural language", i.e., an innate, unlearned language, independent of environmental influences, as part of their arguments. Covarrubias, highly critical of any attempt to attribute such qualities to national languages, states in his entry "Lengua", that "no ay lengua que se pueda llamar natural". In support of his claim, he cites an ancient anecdote, narrated, inter alios, by Herodotus, and which Covarrubias declares to have taken from Juan de Pineda's Monarchia Ecclesiastica. 26 In the narrative, an Egyptian king wishing to determine the identity of the natural and most ancient language, decided to conduct an experiment, having new-born infants raised in a linguistically-deprived environment in order to observe which language they end up speaking. According to Herodotus, the children, reaching the age of two, uttered the word bekos ("becus" in Pineda's version; probably following Valla's Latin translation of the Histories, which reads "beccus"), 27 which turned out to be the Phrygian lexeme for "bread". This lead the king to acknowledge that the Phrygians were the oldest people in the world.
While Herodotus maintains a neutral stance regarding the experiment and its results, Pineda, on whose rendering of the tale (and take on the problem of natural language) Covarrubias relies, rules out the possibility of infantile utterances being motivated. Pineda comments that "de lo de becus de los niños, me parece deverse dezir que lloravan de hambre, y que hazían pucherillos (como dizen) o, si queréis, que oyendo el bec de las cabras le hayan deprendido", reminding readers that the infants in the narrative, albeit unexposed to human speech, may have been influenced by sonic stimuli in their environment. As a way of highlighting the arbitrary assignation of meaning to the mysterious becus, he adds that "sea ansí quanto más que […] en Francia solía sinificar becus el pico de gallo", referring to the Latin lexeme and its use in ancient Gaul. 28 Covarrubias takes Pineda's interpretation one step further by deeming the king's conclusion "cosa vana y ridícula", and reshapes the whole narrative by turning the bleating from a possible explanation to a certain cause, adjusting the infants' utterance, as if wishing to emphasize its onomatopoeic quality: "por aver oýdo los niños el balido de las ovejas, pronunciaron esta palabra bec, que en lengua frigia vale tanto como pan". Having changed beccus (or becus) into bec, the lexicographer can now present the alternative-equally absurd-imposition of meaning more smoothly, by simply mentioning that "también en lengua francesa bec quiere dezir el pico de ave", referring to a contemporary vernacular example, which his readers are more likely to recognize than Phrygian.
Covarrubias endorses here the tone and mode of reasoning of a natural philosopher-if not of a proto-modern psycholinguist-and adds to the discussion a physiological factor, urging readers to consider not only the environment's influence on speech production, but that of the articulatory system. He concludes that in the absence of external stimuli, infants' first syllables would be those easiest to pronounce. 29 And yet, in the middle of this scientifically-structured argument, Covarrubias makes a short, parenthetical remark, according to which "beg o bag"-note the slight change in the vowel and final consonant-"en hebreo vale cibus, pan o otro manjar", accompanied by a representation of the word in Hebrew characters, in transcription, and 29 For other readings of this entry, see Calero Vaquera, "Apud grammaticos", pp. 167-168; Perea Siller, "Historia lingüística universal", pp. 234-235. in Latin translation. Covarrubias never explains the relevance of this information-which not only complicates the reading of his entry but challenges its very argument-for the discussion of natural language, nor does he refer to it elsewhere in his Tesoro. And yet, its graphical presence is hard to ignore.
While Covarrubias's choice of the Hebrew bag is clearly motivated by its resemblance to the sound of bleating or babbling, it is worth noting that, lexically speaking, this is a rather peculiar choice. Bag, a Persian loanword lacking a triconsonantal root or Hebrew cognates, occurs merely six times in the entire biblical corpus, all of which in the Book of Daniel, and is therefore associated almost exclusively with the prophet's story. 30 The narrative, set in Babylonia after the fall of Jerusalem, brings the story of a group of young, noble, Israelite exilesthe prophet among them-that are sent to the Nebuchadnezzar's palace in order to be trained in "the literature and language of the Chaldeans" and become his servants. The captives are assigned a fine menu of the King's food and wine (Hebrew: ‫ָיו‬ ‫ְתּ‬ ‫ִשׁ‬ ‫מ‬ ‫ֵין‬ ‫ִיּ‬ ‫ּמ‬ ‫ו‬ ‫ְֶך‬ ‫ֶל‬ ‫ַמּ‬ ‫ה‬ ‫ַג‬ ‫ַת-בּ‬ ‫ִפּ‬ ‫,מ‬ Vulgate: de cibis suis et de vino), 31 but Daniel and his companions decide not to be defiled by the provisions and remain protected from its corrupting influence.
As Covarrubias does not explain his choice to provide this ostensibly superfluous information, it is hard to determine to what degree the association of infants' first syllables with a biblical intertext dedicated, inter alia, to the teaching and acquisition of language is intentional. Nevertheless, the assignation of a Hebrew meaning to what is elsewhere declared to have no meaning-accompanied by an adjustment of the signifier bekos (becus) to bec and ultimately bag-epitomizes the special status of the sacred tongue in the Tesoro. Situated above and beyond the rules Covarrubias aspires to establish when reasoning about language, Hebrew is included-imposed-at the cost of distorting an otherwise well-structured argument and to the detriment of the accurate transmission of linguistic material-here, the Phrygian word for "bread", with which few, if any, of Covarrubias's readers are likely to be acquainted. 30 Curiously, according to Gesenius, the Persian loanword bag is, or may be, a remote cognate of the Phrygian bekos (Hebrew and Chaldee Lexicon, s.v

Hebrew: Humankind's Lingua Mater
For Covarrubias, Hebrew is not only the language spontaneously produced in the infancy of human individuals, but the one spoken in the world's primera edad, beginning with the creation of the world and ending with Noah. 32 Pertaining to the Sacred History of the world, rather than to the realm of philological investigation, it is hardly surprising that Covarrubias's history of Hebrew-as narrated in his entry "Lengua"-is based on the two language-related myths from Genesis: the creation myth and that of the confusion of tongues.
Alluding to the first, the lexicographer affirms with the greatest degree of confidence: "lo cierto y sin contradicción es que la primera lengua que se habló en el mundo fue la lengua hebrea, infundida por Dios a nuestro primero padre". Hebrew, he states, was the language in which Adam named all living creatures, and after quoting-in Latin, of course-the corresponding versicle (Genesis 2:19), he explains that "el nombre que a cada uno puso era el propio suyo, según su calidad y naturaleza". That is to say, that while other languages maintain an arbitrary relationship between signifiers and what they are supposed to signify, Adam's was an ontologically privileged language: in the beginning, things were in Hebrew. Hence, as he explains in his entry "Etymología", etymologists, whose task is to unearth the origin of words, should trace the first name of things, i.e., the Hebrew one, in which "está encerrado el ser de la cosa". 33 If the creation myth serves Covarrubias to establish Hebrew's privileged relation to the truth of things, the myth of the confusion of tongues enables him to explain Hebrew's relation to other languages that are of relevance to the etymological endeavor: Desta confusión resultaron las setenta y dos lenguas en que se dividieron, y fue ocasión de que siguiendo cada uno la que le fue infundida o confundida, se dividieron a poblar diversas provincias; y no es de maravillar que en lenguas muy estrañas se hallen algunas palabras que tiren a las hebreas, pues desgajándose della, como de su madre, llevassen algún rastro de su primer origen. ("Lengua") Following the distribution of peoples, however, humanity's only segment to preserve the language of their forefathers-or, in Covaru-32 For the Sacred chronology used by Covarrubias, see s.v "Edad". 33 See also s.v. "Nombre". bias's phrasing, "de nuestros primeros padres"-were Eber and his offspring, the Hebrews, after whom the language was named, and to whom an extensive entry ("Hebreos") is consecrated in the Tesoro. The Hebrews' descendants-Jacob and his progeny, also known as the Israelites-preserved the pristine language throughout their captivity in Egypt, but around the time of the Exodus, they too came in contact with other nations and ended up altering and corrupting the language. Thus, Covarrubias explains, by the times of Christ, Hebrew had changed to such an extent that the descendants of the Israelites-or Jews as they were called at that time-no longer spoke it, having shifted to Syriac instead. Thereafter, Hebrew in its pure form was to be found solely in the books of the Old Testament. 34 Indeed, the immediate and most explicit consequence of Hebrew's gradual "corruption" is its conversion from a spoken language to a purely Scriptural one. But the timing in which this process culminates, "quando Nuestro Redentor vino al mundo", gestures towards an equally important trait of Covarubias's Hebrew: its religious identity. If before Christ, the sacred language was spoken by the Jews, God's chosen people, their fall from divine grace coincides with their dispossession of the language. This fall, to which the lexicographer dedicates an extensive section in his entry "Judío", constitutes the theological basis for the antinomy hebreo/judío present in the Tesoro and prevalent in the cultural imagination of early modern Spain. The former is associated with a venerable biblical past which the Spanish "we" aims to appropriate, whereas the latter evokes an abject image of a deicidal people, whose presence in the Peninsula-openly before the expulsion, or covertly after it-is conceived as a threat to the very integrity of Spain. 35 34 "Heber, descendiente del primogénito de Noé y los de su familia quedaron con la lengua antigua, que procedió de nuestros primeros padres, y de su nombre […] se llamó hebrea, y assí quando Jacob y sus hijos fueron a Egipto, los llamaron por este nombre, "hebreos", y el tiempo que estuvieron en la cautividad conservaron su lengua. Saliendo de Egypto y comunicando con otras naciones, la alteraron y corrompieron, de modo que quando Nuestro Redentor vino al mundo, se hablava vulgarmente la lengua syríaca mezclada, y solo los Libros de Moysén y los Profetas conservaron la Escritura y la pureza de la lengua hebrea." ("Lengua") 35 On the antinomy hebreo/judío in Golden Age Spanish drama and literature, see Reyre, Lo hebreo; Fine, "Lo hebreo, lo judío y lo converso"; "El entrecruzamiento".

Arabic's Dubious Lineage
The disappearance of Hebrew as a spoken language marks not only its separation from the Jews: it is also the moment in which, according to our lexicographer, other Semitic languages come into being-"Y assí la caldea como la syríaca se derivaron de la lengua hebrea"-and, more important for our discussion, the moment in which Arabic makes its first, hesitant appearance in Covarrubias's history of Language. It does so, however, in a laconic comment, "y muchos son de opinión que la lengua arábiga trae origen destas tres", which confines its lineage to the realm of the uncertain, without specifying neither who holds such an opinion, nor where the lexicographer stands in relation to them. 36 Curiously, the Suplemento-redacted around the same time the Tesoro was published-brings an alternative theory regarding the origins of Arabic, based on Manuel Correa de Montenegro's Historia de los reyes, señorías y emperadores de España (Salamanca, 1602 and 1608). Correa does not present Arabic as a secondary or tertiary product of Hebrew's corruption, but as one of the seventy-two languages created in Babel, i.e., older and of less obscure or hybrid origins. While the historian expresses doubts regarding the identity of each one of the post-Babelic languages-and their respective bearers-he declares: …tengo por muy probable que las 4 lenguas cardinales que tenemos: chaldea, arábiga, griega y latina, que en el número de las 72, entonces devían ser, como aún oy en día son, las mejores y más universales (quedando reservada la lengua hebrayca para la Iglesia Cathólica y pueblo escogido de Dios) fueron atribuýdas a quatro caudillos supremos, conviene saber, Noé y sus hijos: Sem, Cham y Iapheth. 37 Covarrubias does not comment on this theory, which raises Arabic to the level of a "cardinal" language, ancient and prestigious as Greek and Latin. Paraphrasing Correa's words, however, he modifies the tetrad of principal languages, replacing Chaldean with Hebrew and removing, accordingly, the parenthetical comment on Hebrew being a "reserved" language. Moreover, he eliminates the epithets "cardinales", "mejores", and "universales", omits the statement relating these past 36 As Covarrubias explains in the entry "Opinión", "Distinguen los filósofos la opinión de la ciencia, porque la ciencia dize cosa cierta y indubitable, y la opinión es de cosa incierta. Esta es la causa de aver opiniones contrarias en una misma cosa". 37 Correa, Historia de los reyes, Salamanca, 1608, n.p., my italics. I have not been able to consult the 1602 princeps. qualities to the present ("devían ser, como aún oy en día son"), and offers a lessened version of Correa's original thesis: "Correa […] es de opinión que las cuatro principales lenguas después del diluvio entre las setenta y dos fueron la hebrea, arábiga, griega y latina", confining, once again, these ideas to the realm of opinion, and leaving his own stance on the matter unpronounced. 38 However, the most intriguing silence in the Tesoro regarding the origin, dissemination, or status of Arabic has to do with the theory espoused by Diego de Guadix, whose Recopilación is one of Covarrubias's main sources of Arabic etymologies. In a unique take on the common pan-Hebraist theories of the time, the Granadan friar argues that "la lengua arábiga gana en antigüedad a las demás lenguas del mundo, porque es la lengua hebrea, aunque corrupta", assigning Arabic qualities very similar to the ones Covarrubias assigns to Hebrew. 39 Hence, the presence of Arabic etymons in Spanish and other languages does not stem, as one would imagine, from the expansion of Islam, the presence of Moors in different parts of the world, or contact with them, but from the existence of an ancient linguistic substratum, passed on to the present world by the ancient Arabs, a nation which, according to Guadix, had very little, if anything, to do with Islam or Moors. "[E]stos árabes", he explains with a circular logic, "con esta su lengua arábiga y esta su gramatiquilla arábiga, devieron de cundir o henchir el mundo, porque su lengua arábiga la hallamos derramada y estendida por todo él", making Arabic not only the oldest language, but also a universal one. 40 Hence, even words that the non-Arabist would derive from Latin or Greek, are, in fact, Arabic, as one can learn, for example, from Guadix's etymology of the Spanish hablar, deriving not from the Latin fabulari, but from the Arabic "hebel", meaning "rope", 41 or the Spanish tema, whose origin is not the Greek théma, but rather the Arabic "tamaả, que en arábigo sinifica pretender". 42 Even words of evident 38 Covarrubias, Suplemento, s.v. "Lengua", pp. 332-333, my italics. See also Girón Alconchel, "Covarrubias y la lengua primitiva", pp. 395-398. 39 Guadix,Recopilación,p. 149. For Guadix's theoretical considerations, see García-Arenal and Rodríguez Mediano, "Sacred History, Sacred Languages", esp. pp. 153-157. 40 Guadix, Recopilación, p. 260. 41 Guadix is referring to the Andalusi ḥabal or one of its variants meaning "rope" (Corriente,Dictionary,s.v. ḥbl,p. 114;cf. Classical ḥabl). 42 Guadix is referring to the verb ṭamaʿ "to coveat" (Corriente,Dictionary,s.v. ṭmʿ,p. 334,cf. Classical ṭamiʿa).
Amerindian origin, such as Perú or cacique, are presented as deriving from Arabic, Perú from "berr, que en arábigo sinifica campo o desierto", and cacique from "caciç, que sinifica fraile o religioso", for the indigenous ruler, he explains, "juntamente con enseñorar la república, devía de enseñar religión o buenas costumbres". 43 While Guadix's pan-Arabistic theory underlies most of the entries in the Recopilación, its most articulate expression is found in his preliminary advertencias and in the entry "Algarabía", two sections to which Covarrubias makes no explicit reference in the Tesoro. The uncertainty regarding the exact version of the Recopilación used by the lexicographer makes it difficult to determine whether his silence should be interpreted as an intentional gesture of rejection, or as the result of his lack of access to certain passages in Guadix's text. Given, however, the evidence of other material Covarrubias borrowed the advertencias, 44 one may cautiously suggest that the text consulted by lexicographer was not limited to Guadix's etymologies and that Covarrubias-whose views on Arabic, as we shall see, were radically different-chose to ignore his informant's pan-Arabism altogether, rather than entering an unnecessary polemic.
Be that as it may, Covarrubias's reluctance to elaborate on Arabic's distant history is complemented by his silence regarding the most recent past-and present-of the language in Spain, namely, its eradication and the expulsion of its last native speakers from Iberian soil, starting in 1609. 45 43 "Es caciq, que, en arábigo, significa tu religioso o tu frayle. Advierta el letor que fuere arábigo, qu'el nombre es caciç, que significa frayle o religioso, y la q., es afixo de segunda persona de singular y significa tuyo; assí que / caciq significará tu frayle o tu religioso; y, corrompido, dizen cacique. Devieron de llamar assí a el principal o señor del pueblo, porque juntamente con enseñorar la república, devía de enseñar religión o buenas costumbres (p. 461)". Guadix refers here to qasīs, a dialectal equivalent to the Classical qissīs, "priest, clergyman, pastor" (Corriente,Dictionary,s.v. qss,p. 427). It is, however, noteworthy, that the addition of a possessive pronoun would render the form qasīsak, rather that qasīk (caciq) as stated by Guadix. 44 See Covarrubias's discussion of "sun" and "moon letters", pp. 504-506. 45 A single exception to Covarrubias's silence regarding the explusion of the moriscos is found in the Suplemento, as part of his additions to the entry "Judío". While the parallels the lexicographer draws between the two explusions are of great interest, language plays no part in his text. For a careful, contextualized reading of the passage, see Dopico Black, "Sueños de la nación", pp. 283-294. On the eradication of Arabic and the attempts to resist it, see García-Arenal and Rodríguez Mediano, The Orient in Spain.

The Semitic Presence in Spanish Vocabulary
The contrast between the detailed narrative consecrated to the lingua mater and the laconic reference to its obscure, "corrupted" offspring epitomizes the dissimilar places they occupy in Covarrubias's ideological universe. But while Arabic is accessory, at best, to the lexicographer's universal history of Language, the role it plays in his diachronic account of the development of Spanish far from being marginal. Opening his narrative with an affirmation of Spanish's Romance identity and a cautious expression of doubt regarding the identity of the language spoken in pre-Roman Iberia, Covarrubias explains that Spanish is a "corrupted" form of the Latin once spoken in the Peninsula, whose vocabulary was altered by the vulgo, and modified through contact with other groups and languages. Thus, as one can learn from the entry "Romance", pre-Roman, Gothic, and Arabic words all entered Spanish in different epochs, but while the lexical influence of the pre-Roman and Gothic languages is referred to in neutral terms-the first "blending into" Romance vocabulary, the second "added to" it-Arabic is described as having a more radical, distorting effect, comparable to-or at least associable with-the political circumstances that brought Spanish speakers in contact with them: "y después de la destruyción de España, lo turbaron todo los árabes".
For the purposes of this essay, two aspects of this description should be highlighted. The first is the lexicographer's reference to the quantity of Arabic lexemes entering Spanish ("muchos"), absent in his description of words originating in other languages. The second has to do with his allusion to Hebrew, here not as humankind's primitive tongue, but as yet another national language with which Spanish speakers, "nosotros", had been in contact. This "proximate" facet of Hebrew, which Covarrubias presents without specifying whether it was written or spoken, differs from the ancient, ontologically-privileged language not only in the space it occupies in his metalinguistic narrative, but in the presumed nature of its influence on Spanish: while Adam's Hebrew related directly to the essence of things, this "proximate" Hebrew is mentioned solely in the context of the transmission of words, occurring through contact between groups or individuals ("nos comunicaron vocablos"). 47

From Theory to Practice
In recent years, an increasing amount of scholarly attention has been dedicated to Covarrubias's Hebrew and Arabic etymologies, examining them-separately, for the most part-in light of his ideological postulates, sources, or presumed knowledge. Reyre, whose research focused on Hebrew's privileged status in the Tesoro, explored the lexicographer's attempts to "arrimar el hebreo a su castellano", as she puts it, and, in particular, to subject Spanish words to Hebrew morphological and phonetic principles. Among the salient manifestations of these attempts, she highlights the derivation of Castilian verbs from Hebrew triconsonantal roots; the attribution of "natural" qualities to Hebrew sounds; and the treatment of Hebrew orthographical signs as bearing transcendent meanings-all showcasing Covarrubias's biblical and grammatical learning. 48 My own disquisitions, following Reyre's, focused on the tension between Covarrubias's concept of Hebrew and moments in which he shows himself incapable or unwilling to employ the very grammatical rules he pretends to appropriate. The attempts to establish an affinity between Spanish and Hebrew, I argued, entailed not only adjustments of Spanish to Hebrew logic, but also a conscious adaptation of the lingua mater's sounds, forms, and meanings in order to make it resemble its alleged offspring. 49 In a similar vein, though more systematically and on a significantly broader scale, György Sajó's examination of Covarrubias's entire Hebrew corpus offered a taxonomy of the explicit mechanisms employed by the lexicographer in order to explain the evolution of Hebrew etymons into Spanish lexemes, and, no less important, the concealed methods he uses in order to achieve this goal, including manipulative transcriptions, semantic adaptations of Hebrew words, the invention of Hebrew pseudo-lexemes, or the creation of impossible etymons through grammatical manipulations or combinations of Hebrew, Arabic, and Aramaic. 50 While these and other readings differ in the degree of intentionality and meaning they attribute to the lexicographer's idiosyncrasies and deviations from the rules of Hebrew grammar, it is generally agreed that his Hebrew etymologies-fabricated, in their vast majority, by Covarrubias himself-are to be interpreted in light of his metalinguistic narrative and declared agenda of appropriating the holy tongue. Drawing on a narrative that situates the sacred language in a privileged ontological realm, Covarrubias establishes connections between Spanish and Hebrew lexemes without adhering to the same kind of linguistic reasoning required in other cases, and without committing-for the most part-to historical contact between groups of speakers as the basis for lexical borrowing. Moreover, the lexicographer's choice to stick to an ancient, "uncontaminated" register of Hebrew-with only a handful exceptions referring to Hebrew in a post-biblical context 51 -not only 49 Hasson, "La lengua santa". 50 Sajó, "Las etimologías (i)", "Las etimologías (ii)". 51 Only a handful of entries in the Tesoro mention Hebrew in an explicitly post-biblical context or associate it with the historical presence of the Jews in Spain (e.g., "Barahá", where Hebrew is mentioned in a Jewish liturgical context; "Godo", where the lexicographer distinguishes the biblical sense of the word gôy, "gens, natio", from its Jewish use as a term referring to non-Jews; or a few place names around Toledo, attributed to the Jews enables him to underplay the language's connection to the Jews in his narrative, but technically facilitates his etymological endeavors by limiting the search of etymons to a finite, well-charted, biblical territory, accessible even to a second-class Hebraist like Covarrubias by means of grammar books and lexicons. As for the Arabic etymologies, motivated by the need to address a real, proximate presence in Spanish vocabulary-which, paradoxically, is less accessible to the lexicographer than ancient Hebrew-they pose a different set of questions and methodological challenges. Unlike Hebrew, Arabic is identified in the Tesoro with a concrete group-or groups-of speakers, whose customs, traditions, and histories often emerge from the discussion of Arabisms. Thus, the diverse linguistic material archived in the Tesoro-ranging from classical etymons presented with case-endings to colloquial and markedly dialectal formsis accompanied by a rich gallery of cultural representations. While the provenance, accuracy, and, most important, the degree of "Spanishness" attributed to these representations still awaits a meticulous evaluation, their mere presence suggests that in spite of-or alongside-the lexicographer's unequivocal stand on religious matters and adherence to the discursive paradigm of the Reconquista, he conceived Andalusi, morisco, and Moorish realities as part of local history, worthy, at least in principle, of being part of the discussion. 52 Considering Covarrubias's lack of first-hand knowledge of Arabic and consequent reliance on the work of others, the Arabic entries in the Tesoro pose, first and foremost, the challenge of discerning the lexicographer's voice from that of his informants and determining the degree of intentionality ascribable to the rich linguistic material they provide. How much of the linguistic material is Covarrubias capable of-or interested in-understanding? To what extent is he aware, for who once inhabited them). Perea Siller, who links Covarrubias's philohebraism to his converso descent, reads these toponymic manifestations of "proximate" Hebrew as a natural continuation of the lexicographer's pan-Hebraist theory ("Historia lingüística universal", esp. pp. 242-244). Weighing, however, these isolated cases against Covarrubias's explicit anti-Jewish and anti-converso discourse, and no less important, against hundreds of other etymologies reflecting the lexicographer's ahistorical (or sacrohistorical) approach to Hebrew, I deem them exceptions proving the rule. See Hasson, "La lengua santa", 269-272; Sajó, "Las etimologías (I)", pp. 139-141. 52 For a discussion of encyclopedic discourse in the Tesoro, see, Neyrod, "L'héritage arabe", pp. 109-110; Calderón, "El Tesoro de la mora", pp. 88-91.

example, of the diglossic condition of Arabic manifest in his Tesoro?
And finally, what role does his knowledge of Hebrew play in his attempts to understand or explain Arabic etymologies?
One approach to the problem is that of Stefan Ruhstaller, whose painstaking study of Covarrubias's Arabic etymologies vis-à-vis his sources emphasizes the lexicographer's role as a recopilador, who, in spite of his shortcomings, selectively incorporates valuable material into his Tesoro, recognizing the structural and theoretical superiority of Urrea's explanations; excluding most of Guadix's etymologies; and further developing Arabic etymologies towards their presumed Hebrew source. Given Covarrubias's philohebraism, Ruhstaller explains, this latter practice does not imply "ninguna crítica o enmienda" of the Arabic sources, since the Hebrew etymology is conceived as the origin of the Arabic, rather than an alternative to it. 53 A different path is the one taken by Dominique Neyrod, who in a series of compelling studies on the Arabic etymologies that Covarrubias attributes to Urrea, highlights a notable and yet unexplored grammatical discourse. In various cases, she convincingly demonstrates that the discourse consists of semantic calques echoing Arabic grammatical traditions. Additionally, Neyrod shows some of Covarrubias's attempts to participate in the discussion of Arabisms in a variety of ways, e.g., the addition of intermediary forms between the Arabic etymon and the Spanish lexeme; phonetic observations; and the illumination the Arabic grammatical phenomena through allusions to parallel phenomena in Hebrew. Based on these interventions, Neyrod goes as far as questioning Covarrubias's preliminary declaration of ignorance-which she deems "politiquement correcte"-and suggests that it has more to do with the general climate of prohibitions on Arabic in Covarrubias's Spain than with his actual knowledge of the language. 54
Given, however, that the identification of equivalent roots depends, first and foremost, on the ability to recognize the consonants comprising them, Covarrubias's method reveals, in numerous occasions, his inability to identify Arabic consonants, and in other occasions, his lack of interest in doing so. This, for example, is what occurs in the case of the Arabic word for "fortress", qalʿa, which, given its remarkable presence in Iberian toponymy, is discussed in various entries ("Alcalá", "Calatayud", "Calatañoz") as well as an independent lexical unit ("Cala"). "Cala", Covarrubias states, "significa alguna vez villa, castillo o fortaleza, y entonces es nombre arábigo de raýz hebrea ‫ָא‬ ‫ָל‬ ‫,כ‬ cala, prohibere". The lexicographer associates the Arabic noun and the Hebrew verb by explaining that a fortress "prohibe a los enemigos entrar dentro", and brings, in support of his claim, the etymology of the Latin arx, "fortress", which, connected to the verb arceo, "to confine, prohibit access", follows a similar logic ("como se dixo arx, ab arcendo, porque los arriedra"). What he fails to recognize, however, is that the Arabic etymon, qalʿa, is spelled and pronounced with a qāfequivalent to the Hebrew qôf rather that the kāf in kala-and, more importantly, with an ʿayn, equivalent to the Hebrew ʿayin, to whose pronunciation the lexicographer dedicates special attention elsewhere the Tesoro. 64 One could attribute this inaccuracy to local traditions preserving the semantic, but not the phonetic qualities of the Arabic qalʿa-or to the fact that Guadix-on whom Covarrubias relies, at least in part in this etymology-transcribes the etymon without marking the ʿayn, in spite of his general tendency to do so in other cases. 65 The entry "Calatayud", however, suggests that Covarrubias also consulted this etymology with Urrea, who was highly unlikely to misspell or mispronounce such a common word: El nombre es arábigo, Cala, y Taud […]. Dize Diego de Urrea que Calatayud vale "castillo" o "villa de Iob", […] y dizen averla fundado Aiub, que vale Iob, moro governador de la España después que la ganaron los alárabes […]. 64 See s.v. "Arraax" (quoted in note 70 below), "Gangoso", and "Guaya", all mentioned in Reyre,"Cuando Covarrubias",p. 16. 65 Throughout his references to toponyms containing "cala", Guadix spells the etymon without indicating the ʿayn, and defines it, almost with no excpetion as "castillo, i., que haze frontera", a definition reproduced, inter alia, in Covarrubias's entry "Alcalá". See, e.g., Guadix,Recopilación,s.v. "Alcalá",p. 227; "Calatayud", p. 477.
Since we do not know whether Covarrubias received Urrea's explanation in a written or oral form, it is hard to determine whether he heard the royal interpreter pronounce the word or merely saw it in its transcribed form. What we can learn from the Arabist's translation, "Castillo o villa de Iob", is that Urrea had in mind the construct state Qalʿat Ayyūb, in which the tā' marbūṭa of qalʿa is realized like a tāʾ, marking thereby its relation to its founder. 66 While Covarrubias is capable of associating the Arabic name Ayyūb with the biblical Job, either based on Urrea's explanation or drawing on the resemblance of Arabic Ayyūb and Hebrew Iyyôḇ, his segmentation of the toponym into "cala, y ta[i]ud" reveals how far he is from knowing the derivational scheme of the word, or from identifying a common, basic structure, whose Hebrew equivalent is clearly explained in his sources. 67 In any case, the only thing one can assure is that the Arabic ‫قلعة‬ and the Hebrew ‫כלא‬ coincide only in this particular transcription.
Another example in which Covarrubias incorrectly determines the radical consonants of an Arabic lexeme based on his knowledge of Hebrew can be found in the entry "Alfahar" (from fakhkhār, "earthenware"), where he presents the etymology, without mentioning his source, as "fahar, que sinifica barro, pero su raíz de hebrea, ‫,עפר‬ hhafar, que vale tierra", explaining the difference between the immediate and the ultimate etymons as the result of a metathesis: "el arábigo trocó las sílavas, y dixo fahar por hafar". 68 While metathesis is not, in itself, foreign to Semitic linguistics, the lexicographer's explanation assumes that the phonetic value of both h and hh in these etymons is identical, while in Arabic it stands for khāʾ, and in Hebrew, ʿayin. 69 66 Cf. Guadix's segmentation: "Consta de calat, que en arábigo significa castillo, i., que haze frontera, y de yub, que significa Job" (Recopilación,s.v. "Calatayud",p. 477). 67 See, e.g., the grammatical summary in Sante Pagnini's Thesauri Epitome: "Feminina in He: ‫,ה‬ in regimine singulari et affixis, ‫ה‬ perpetuo conuertunt in ‫".ת‬ [Feminine nouns ending with a he: when in construct states or when an affix is added to them, the he always turns into a tav] ("De nominibus femininis", n.p.). 68 The princeps reads ‫צפר‬ [ṣafar] instead or ‫ﬠפר‬ [ʿafar]. Given the transcription and translation that follow, this typographical error seems to be the typesetter's, rather than the lexicographer's. 69 Another example can be seen in "Aljamía", where the Arabic root jmʿ is associated with the Hebrew yam.

Lost in Transcription
Such attempts to infer the roots of Arabic words-i.e., their Hebrew roots-can, of course, be explained by Covarrubias's declared ignorance of Arabic, and, more concretely, of its alphabet. But ignorance, like knowledge, comes in many forms, and one can be completely or partly unfamiliar with a language's alphabet, and yet conscious of what transcription can convey and what it cannot. Based on hundreds of transcriptions of Hebrew in the Tesoro-which, as Sajó shows, come from a variety of systems Covarrubias adopts according to need-it can be affirmed beyond doubt that the lexicographer was aware of the fact that the Latin characters never replace Hebrew letters, but represent the approximate sounds of phonemes, and, no less important, that they are far from conveying unequivocally Hebrew consonants. 70 Is this also the case with Arabic letters and consonants? Covarrubias's does not offer an explicit, systematic discourse revealing how much of the Arabic alphabet he knows, nor to what extent he imagines Arabic consonants to resemble or differ from Hebrew ones. One encounters sporadic references to three Arabic letters-alif, mīm, wāw-brought alongside their Hebrew equivalents-aleph, mem, vav 71 -and even one case in which a presumably Arabic lexeme is represented in Hebrew characters, 72 which could convey the impression that even if Covarrubias was unfamiliar with the graphic representation of Arabic letters, he was aware of the consonantal value of some of them, and, to some extent, of their changing qualities in different derivational schemes.
Had the lexicographer used this information, he could have easily associated it with the biblical ‫,בצל‬ "onion", mentioned by Sante Pagnini. words in which the lām of the definite article is pronounced (e.g., Alcalá) and ones in which it is assimilated to the initial consonant (e.g., azúcar) was of great relevance for the analysis of Arabisms, a fact of which Covarrubias was well aware. Thus, we find two different explanations of this phenomenon in the Tesoro. The first, framed as "dotrina de Diego de Urrea", presents the general rule of assimilation without specifying which letters count as "sun letters" and which as "moon letters". 73 The second, brought with no explicit reference to a source, is based on Guadix's introductory advertencias, where the Granadan friar explains that the Arabic lām, …aunque escri[t]a en la dición, no se a de leer, ni a de sonar la pronunciaçión quando se le siguiese alguna letra de las que los árabes llaman gemcías, i., solares, que son las sig [u]ientes: ‫ن‬ ‫ظ‬ ‫ط‬ ‫ش‬ ‫س‬ ‫ذ‬ ‫د‬ ‫ز‬ ‫ر‬ ‫ث‬ ‫ت‬ ‫ص‬ ‫,ض‬ el sonido de las quales corresponde a el sonido de nuestra t., r., z., d., c., x., n….". 74 While Guadix's explanation is wanting in many senses-he neglects to mention, for example, that the disappearance of the l is compensated by a doubling of the following consonant, or that lām, too, falls into the category of ḥurūf shamsiyya-it conveys a clear awareness of the limitations of transcription: the seven Latin letters in his list are not presented as equivalent to the thirteen Arabic letters, but as representing their sounds. 75 Covarrubias incorporates this information into the entry "Al", dedicated to the Arabic article, and attempting to integrate Guadix's explanation with what he grasped from Urrea's, writes: 73 "Los árabes usaron del artículo el para todos los géneros de nombres, y nosotros le trocamos en al, y denota nombre apelativo y común, porque los propios no tienen artículo, por ser de suyo determinados. Algunas vezes del al se pierde la l, quando los nombres empieçan en letra solar (que llaman los arábigos) y entonces la letra iniciativa de la tal dicción suena dúplice, etc. Esta es dotrina de Diego de Urrea." (s.v. "El") Note that while the rule of assimilation in the second sentence is explicitly attributed to Urrea, it is difficult to discern the informant's voice from that of the lexicographer in the commentary on the supposed evolution of el into al, which establishes a distinction between "los árabes" and "nosotros".
Al es ni más ni menos artículo arábigo que llaman solar, y quando la dición a quien se ajunta es de las que los arábigos llaman gemeia [sic], pierde la l y pronúnciase tan solamente la a, y las letras son las que corresponden a estas que siguen: t, r, z, d, c, x, n. While Covarrubias's explanation is offered with a gesture conveying certainty and precision ("ni más ni menos"), the hybrid technicism he coins, "artículo solar" (as opposed to Urrea's calque, "letra solar"), betrays his lack of distinction between the article and its phonetic realization. This confusion is complemented in other entries, where the equivalent "artículo lunar", is employed, ironically, where the Arabic article is followed by a sun letter. 76 Be that as it may, Covarrubias's reproduction of Guadix's list-omitting the Arabic characters but preserving the Latin ones in the same order they are found in the Recopilación-reveals an essential trait of his concept of transcription when it gets to Arabic: the Latin letters no longer correspond to the sounds of the absent Arabic letters represent, but to the letters themselves.
This being the case, one understands why Arabic etymons such as those of cala or alfahar can be assigned Hebrew roots that resemble them only in their Hispanicized form. In these and other cases, the lexicographer bases his etymological inferences on transcriptions, without asking himself-or Urrea-which Arabic consonants they represent. An extreme example of this reliance on transcription, although unrelated to Hebrew, can be found in the etymology of the Valencian toponym Cirat-from the Arabic ṣirāṭ-which Covarrubias attributes to Guadix: "Este nombre Cirat, leyéndole al revés dize Taric, o Tarec, que vale camino". As the Recopilación does not consecrate an entry to this particular toponym, it is difficult to determine whether this is an original contribution of Covarrubias, a popular etymology he may have picked up in his Valencian years, or an observation found in another source, authored by Guadix or by someone else. In any case, the toponym's semi-palindromic quality exists exclusively in this particular Hispanicized form, in which the letter c can represent both the ṣād in ṣirāṭ and the qāf in ṭarīq. 76 Tesoro, s.v., "Acelga", "Adarme".

Hebrew as an Alternative to Arabic
In other entries, however, Covarrubias's Hebrew etymologies of Arabisms not only cannot be read as attempts to understand or clarify the Arabic, but either contradict the morphological or phonetic explanations offered by Urrea, or derive Arabic lexemes from Hebrew ones without trying to pass through a supposedly shared root. Let us examine, for example, what occurs in the entry "Almohada": …dize Diego de Urrea que en su terminación arábiga se dize mehaddetum, del nombre haddũ, que sinifica mexilla, y por ser nombre local almohada, tiene la letra m, o la partícula mo, que sinifica lugar, cosa sobre que está otra, y assí al-mo-haddetũ corrompido dezimos almohada. Sin embargo desto, digo que puede ser nombre hebreo, del verbo ‫ַד‬ ‫ָע‬ ‫,מ‬ mahad, que sinifica declinare, reclinare, y sobre el almohada reclinamos la cabeça. This Arabic etymology, thoroughly explicated by Urrea and slightly modified by Covarrubias's inappropriate addition of the tanwīn to a definite noun in al-mo-haddetũ, contains all the morphological and semantic information necessary to infer the radical consonants, or-assuming one is incapable of identifying the consonant represented by the h-at least discerning the elements in the transcription that form part of the root and those that do not ("la letra m, o la partícula mo"). The Hebrew etymology, in contrast, not only turns the m of almohada into a radical consonant, but adds a semantic adjustment of the verb maʿad, which means "to fall", rather than "to decline". 77 Like "Almohada", the entry "Alguazil" offers a Hebrew etymology that contradicts Covarrubias's Arabic sources, and, in particular, goes against the detailed phonetic explanation provided by Urrea. The entry, which opens with a quotation from the Siete Partidas affirming the Arabic origin of the term ("Alguazil llaman en arábigo aquel que ha de prender e de justiciar los omes en la Corte…"), is followed by two alternative Arabic etymologies: the first, attributed to Guadix, is "de al-guazir, que vale ministro de justicia". 78 The second is offered by Urrea, who, unaware, perhaps, of the alternance of r and l, suggests an etymon from the root ‫,وصل‬ which resembles more the form alguazil-"en su terminación se dize vesilun, del verbo vesale, que sinifica allegar, y assí alguazil será allegador y recogedor, no digo de haziendas, sino de delinqüentes…". This semantic clarification is followed by a detailed phonetic observation, which, albeit attributed to Urrea, seems to integrate the lexicographer's own comparative perspective, as one can learn from the reference to Hebrew: Y porque el sonido desta etimología no quadrará a todos, advierte el dicho Urrea que todos los nombre[s] que empieçan [con] gua, los árabes los pronuncian por va, con la ‫וּ‬ vau de los hebreos, pero los moriscos de España la pronuncian gua, de modo que por al-vasil, dizen al-guazil.
The concept of Arabic emerging from the reference to different pronunciations among árabes and moriscos leaves much room to debate: one could, along with Neyrod's observations, read the quoted passage as reflecting the lexicographer's adherence to a historical framework, in which the trajectory of Arabisms is delineated from their presumed classical origins, passing through dialectal variants, and all the way to their Hispanicized form. 79 Conversely, one could read it as an observation on regional differences, which Urrea may have ascribed to a historical or dialectological perspective, but which, as far as Covarrubias allows us to reconstruct his understanding, could have been ahistorical just as well. 80 In any case, the passage establishes beyond any doubt that the root of alguazil begins with the Arabic equivalent of the Hebrew vav-to put it in Covarrubias's terms-and that the g in alguazil reflects the local realization of the letter, rather than the radical consonant itself. And yet, Covarrubias's Hebrew etymology, albeit introduced with a gesture of modesty, dismantles the whole construction: "sin perjuyzio de lo 78 Cf. López de Velasco, Orthographía, "alguacil, guazir" (p. 255). 79 "L'arabe est commenté comme langue historique, ayant évolué d'une forme classique vers une forme régionale et dont de nombreux mots, "corrompidos por el vulgo", sont entrés à part entière dans la langue espagnole." (Monument/Document, p. 68) 80 On a dialectological perspective in Urrea, see Ruhstaller, "Covarrubias como recopilador", p. 321. dicho-porque yo doy gran crédito a Diego de Urrea-podríamos dezir que al-gazil [sic] es hebreo del verbo ‫ָל‬ ָ ‫ָז‬ ‫ג‬ [sic], gaçal, rapere, porque echa mano del delinqüente", suggesting a Hebrew root beginning not with a vav, but with a gimel, and omitting, intentionally or unintentionally, the u of the Spanish alguazil. 81 The Hebrew in these examples functions as an alternative to the Arabic etymologies, contradicting Urrea's morphological explanation in almohada, and his phonetic observation in alguazil, rendering whatever information they could have provided the lexicographer about their Arabic root irrelevant. Other entries indicate that, at times, Covarrubias is indifferent to his declared rootbased method altogether. This can be observed, for example, in the etymology of "Buz", "el beso de reverencia y reconocimiento que da uno a otro", in which the lexicographer affirms: El nombre es arábigo, de nibuz, que vale besar, y buz, beso. Está tomado del hebreo en esta forma, que una de las vocales de los hebreos se llama quibuz, y es la u, tres puntos en escalerilla ⋰ del verbo ‫ּץ‬ ַ ‫ָב‬ ‫,ק‬ Kabbats, 82 id est, congregare, porque para pronunciar la dicha vocal u es necessario congregar y juntar los labios.
Covarrubias does not mention a source for this Arabic etymon, which can be found in this markedly dialectal form in Alcalá's Vocabulista ("besar la muger, nibŭç, becězt, buç"). 83 While the root of this particular verb could have caused certain confusion to readers familiar with the classical bāsa, "to kiss", due to the gemination that occurs in the 81 See Neyrod, "L'héritage árabe", pp 112-113. 82 The vocalization ‫ּץ‬ ַ ‫ָב‬ ‫ק‬ is a non-existent hybrid of qaḇaṣ (in qal) and qibbeṣ (in piʿel), reflected also in the transcription Kabbats. Rather than a typographical error, the superfluous dageš in this case seems to derive from Luis de São Francisco's explanation, "dicitur a verbo ‫ַץ‬ ‫ַב‬ ָ ‫,ק‬ Kabbats, id est, congregare" (Globus, p. 80), which Covarrubias quotes almost verbatim. Given the lexicographer's interest in Hebrew orthographical signs, and particularly the dageš, to which he dedicates an entry in the Tesoro (s.v. "Dages"), his failure to notice the misuse of the dageš here can be regarded as yet another example of the tension between pretense and actual grasp of the language. 83 As Ruhstaller shows, Covarrubias's reliance on material from the Vocabulista is mediated, for the most part, by López de Velasco's Orthographía ("Covarrubias como recopilador", p. 326). Given, however, that this particular lexeme does not appear-to the best of my knowlegde-in the Orthographía, this could be an indication that Covarrubias also consulted the Vocabulista directly.
Andalusi form (bws>bss), 84 any reader following Covarrubias's advice, "no hazer mucho caso a las vocales", could have come up with the letters equivalent to b, u, z-looking at the conjugated present tense-or with b, z, z-looking at the past tense. In fact, one of Covarrrubias's contemporaries, Francisco del Rosal, seems to have done just that in his Origen y Etymología de todos los vocablos Originales de la Lengua Castellana, where the etymology of "Buz" is presented as follows: Buz […] es palabra hebrea y arábiga, y quiere decir abatimiento, menosprecio, humilde reverencia […] de otro verbo buç [on the margin, ‫]בוּז‬ que es estimar en poco y abasallar. De donde el arábigo dice buç [on the margin, ‫]בוּס‬ al besar la muger. 85 Looking at the same entry from Alcalá's Vocabulista (as one can learn from the translation, "besar la muger"), and similarly motivated to find an "ultimate" Hebrew etymology for the Arabism, Rosal comes up with an explanation that, in spite of its semantic shortcomings, is nevertheless an attempt to associate an Arabic word with a Hebrew one-both represented in transcription and in Hebrew charactersbased on their presumably shared or similar root. Covarrubias, in contrast, is invested in explaining the Hebrew name of the vowel u, in its graphic representation (which, in spite of the detailed description, is printed backwards), and in showcasing his Hebrew learning by adding a reference to São Francisco's grammar book. So much so that Covarrubias is completely oblivious of his own declared root-based method.

Conclusion
Examined from a comparative perspective, the representation of Arabic and Hebrew in the Tesoro reveals not only the radically dissimilar roles they play in Covarrubias's ideological world, but the complex, and, at times, contradictory interplay of knowledge and ignorance that shapes his etymological practices. Narrated in detail and with a sense of utmost certainty ("lo cierto y sin contradicción"), Covarrubias's history of Language establishes Hebrew as humankind's lingua mater, a purely Scriptural-and, for the most part, de-Judaized-language, whose knowledge provides access 84 Corriente,Dictionary,s.v. bss, Origen y Etymología, s.v. "Buz", fol. 72r (p. 147 in facsimile).
to the very essence of things. Positioning himself as a possessor of such knowledge-and committed to showing the affinity of Spanish and Hebrew-the lexicographer showcases his learning throughout hundreds of etymologies, favoring this kind of knowledge over the need to produce coherent arguments, and, at times, at the expense of an accurate transmission of linguistic material.
Arabic, in contrast, is characterized first and foremost as a language that is unknown. While the lexicographer acknowledges Arabic's massive influence on Spanish vocabulary-and consequent importance to the etymological endeavor-his ideological and historiographical discourse on it is laconic, and, for the most part, accessory. He is reluctant to pronounce his own views on the language's origins-beyond a general notion of Arabic deriving somehow from Hebrew-and, while recognizing its association with Islam, he refrains from narrating its distant or proximate history.
While the lexicographer declares to have used his knowledge of the holy tongue to validate whatever information his informants provide him on its "corrupted" offspring, this knowledge plays a contradictory role in his treatment of Arabic. Indeed, at times, it can be said to compensate for his lack of acquaintance with Arabic insofar as it allows him-through a comparative method-to illuminate grammatical phenomena that would otherwise seem obscure, or get closer to the meaning of Arabic lexemes by examining their Hebrew cognates, which he deems to be etymons. In other cases, however, this very method reveals the lexicographer's ignorance of Arabic roots, letters, and consonants. While some of these erroneous inferences can be attributed to Covarrubias's lack of access to knowledge or need to rely on partial information, others reflect his refrainment from inquiring into the roots of Arabic words, even when he is provided with sufficient tools to do so. In these cases, the lexicographer's knowledge of Hebrew can be said to promote the ignorance of Arabic, or least obviate knowledge thereof.
Alluding to Quevedo's critique of Covarrubias's methods, one could frame these inconsistencies as yet another manifestation of the lexicographer's "erudición desaliñada", 86 or, assuming a more empathic stance, think of them as the result of an uneven battle between an insufficiently trained etymologist and wild beasts, whose forms, and es-pecially sounds, are too foreign to tame. But such framings neglect to consider the intentional component of the complex dynamics between knowledge and ignorance, resulting from the need to show the resemblance of Spanish with Hebrew.