Moïnfar calculates that the Šāh-nāma contains 706 words of Arabic origin, occurring a total of 8,938 times. The 100 words occurring most frequently account for 60 percent of all occurrences.




It is well known that the Šāh-nāma has comparatively few Arabic loanwords. Most commentators claim casually, like E. G. Browne (1862-1926; q.v.), that Ferdowsi consciously “avoided their use ... because he felt them to be unsuitable to the subject of his poem, but even in his time many Arabic words had become so firmly established in the language that it was impossible to avoid their use.” (II, pp. 145-46). Browne goes on to sample 21 verses from the episode of Sohrāb and Rostam, and estimates the Arabic vocabulary in the Šāh-nāma at four or five percent.

The one study offering a complete count and a comprehensive commentary on the Arabic vocabulary of the Šāh-nāma is the Swiss dissertation by Mohammad Djafar Moïnfar, written under the guidance of Emile Benveniste (1902-76) and Gilbert Lazard (b. 1920). Moïnfar rightly criticizes previous attempts for their incomplete samples and loose criteria, such as lax etymology, inclusion of proper nouns, and counting derivatives, compounds, and components of collocations as separate words (pp. 1-2). His study is based on the 1935 Glossar zu Firdosis Schahname by Fritz Wolff (b. 1880) and supplemented with references to the text of the Šāh-nāma in editions published up to the 1960s. For semantic categorization Moïnfar adapts the system that Rudolf Hallig (1902-64) and Walther von Wartburg (1888-1971) first applied in 1952 to French lexicography.

Moïnfar calculates that the Šāh-nāma contains 706 words of Arabic origin, occurring a total of 8,938 times, which yields 8.8 percent of Arabic in the vocabulary (i.e., individual tokens), and a frequency of occurrence of 2.4 percent (pp. 61, 65). As is usual in this type of study, a relatively small number of words in the inventory accounts for the majority of occurrences in the text. The 100 words occurring most frequently account for 60 percent of all occurrences of Arabic loanwords (pp. 62-63, compare the ranked frequency list pp. 72-87). The most frequent word is ḡam(m) “sorrow, grief,” with 707 occurrences; next is hawā(ʾ) “air,” with 435; and third comes bar(r) “land” with 239, often collocations such as bum o bar “territory.” These loanwords illustrate, incidentally, the typical reduction of a final geminate consonant and deletion of the final hamza when an Arabic word with these consonant combinations is incorporated into Persian. Of least frequency are 258 words occurring once only; these include šarṭ “stipulation,” noqṣān “deficiency,” and baḥr “sea.” Of these words of unique occurrence, 104 are noted as missing or having variants in some texts (pp. 3, 81-84). Other modifications of borrowings include the loss of an initial vowel, as in amir > mir “commander” and abu > bu “father” the substitution of an epenthetic for a prothetic or other initial vowel, as in aṭnāb > ṭanāb “rope” and Aflāṭun > Falāṭun “Plato;” the distribution of words with the Arabic feminine ending as -at or -a (p. 66; see ARABIC ELEMENTS IN PERSIAN, Suppl.); and metathesis, as in ketf > keft “shoulder” (p. 67).Grammatical influence is restricted to the incorporation of some Arabic broken plurals, such as aṭrāf “sides, region” and fonun “ways, arts.”

There are 89 personal and place names in the text, which are not counted for statistical purposes (pp. 4-11, 61). Also excluded are 28 arabicized Persian or Iranian words (pp. 14-18, 61), eight words of probable Sanskrit provenance (pp. 18-19, 61), 30 Greek and Latin loanwords in their Arabic or Persian forms (pp. 19-22), and 21 Iranian loanwords in their Arabic forms (pp. 67-68). Arguably the chess terms roḵ(ḵ) “rook” and māt “mate” (p. 45), which are included, should be relegated to one of the first two categories (see CHESS. ii. CHESS TERMINOLOGY; note that roḵ(ḵ) has no more connection with the homonymous bird than does its English equivalent rook). Of the semantic categories, two of the largest sub-categories, as might be expected, are that of army (pp. 49-51) and warfare (pp. 51-53).

Moïnfar notes that a larger proportion of Arabic vocabulary is found in the Šāh-nāma’s introduction and conclusion, neither of which is integral to the epic’s narrative (p. 62). These passages include Ferdowsi’s eulogy and satire (see FERDOWSI ii. HAJW-NĀMA) of the Ghaznavid sultan Maḥmud (r. 998-1030). Since they treat contemporary matters and have recourse to polite and pious formulae, typically they have a high proportion of unique occurrences: e.g., wojud “existence,” ḥamd “praise,” hijrat “hijra.” The satire (Šāh-nāma, ed. Mohl, I, p. lxxxviii-xcii) is 93 verses long, and has 45 Arabic words, or 48.2 percent.

Ferdowsi himself, though arguably a language nationalist (ʿAjam zenda kardam bedin Pārsi “I have brought the Persians to life with this Persian” in the satire—if this line is authentic, cf. Amin-Reyāḥi, p.28 note 1), does not state explicitly that he minimized his use of Arabic (see FERDOWSI i. LIFE, p. 518). Moïnfar, citing the variable proportions above, takes the view that the constraints of material and style were enough in themselves to reduce the count in more Persian contexts, and concludes that the poet did not discriminate against Arabic loanwords (p. 61). Supporting this position is the fact that many of Ferdowsi’s Arabic words are not neologisms, but synonyms for ordinary Persian words that he uses elsewhere: e.g., Ar. donyā, ʿālam = Per. jahān, giti “world,” Ar. bar(r) = Pers. bum, zamin “land,” Ar. ḡam(m) = Per. anduh or andoh “sorrow,” Ar. tannin = Per. aždahā “dragon.” Like later poets, Ferdowsi used Arabic borrowings for Persian words in elegant variation, or where necessary to fit the meter.

There was a practical constraint on acceptable Arabic vocabulary, because the syllabic structure of Arabic loanwords had to fit the motaqāreb meter of the Šāh-nāma: ᴗ - - / ᴗ - - / ᴗ - - / ᴗ - //. Moïnfar illustrates cases where five loanwords were altered to fit the meter (p. 67). It should be noted, however, that it would be impossible to accommodate any words of the patterns motafāʿel, motafaʿʿel, estefʿāl, or mofāʿala(t). Verbal nouns (maṣdars) of roots with a final weak consonant that are formed on this last pattern may, and do, occur: e.g., moḥābā “respect,” modārā “moderation,” mokāfāt “recompense.”



Moḥammad Amin-Riāḥi, Sar-čašmahā-ye Ferdowsi-šenāsi: Majmuʿa-ye neveštehā-ye kohan dar-bāra-ye Ferdowsi o Šāh-nāma o-naqd-e ānha, Tehran, 1993.

E. G. Browne, A Literary History of Persia, 4 vols., Cambridge, 1902-24.

Abu’l-Qāsem Ferdowsi, Le livre des rois, ed. and tr. Julius Mohl, 7 vols., Paris, 1838-78.

Rudolf Hallig and Walther von Wartburg, Begriffssystem als Grundlage für die Lexikographie: Versuch eines Ordnungsschemas, Deutsche Akademie der Wissenschaften zu Berlin: Veröffentlichunges des Instituts für Romanische Sprachwissenschaft 19, 2nd ed., Berlin, 1963.

Mohammad Djafar Moïnfar, Le vocabulaire arabe dans le Livre des rois de Firdausī: Étude philologique et de statistique linguistique, Wiesbaden, 1970; originally, Ph.D. diss., Université de Neuchâtel, 1968.

Fritz Wolff, Glossar zu Firdosis Schahname, Berlin, 1935; reprint, Hildesheim, 1965.

(John Perry)

Originally Published: July 20, 2005

Last Updated: June 23, 2010