¹ Compare: L'Ottimo: “il viso verso l'anche e verso le gambe di Lucifero rivolse . . . ”; Bambaglioli: “Virg. volvit faciem versus ancas et tibias Luciferi.. . ”; Daniello: “volse la testa ove (Lucifero) avea le zanche . . . ”; Garboli: “elli: Lucifero”; Chimenz: “Virgilio . . . volge la testa verso le gambe [zanche] di Lucifero.”

As for the second interpretation, twelve commentators make it clear in some way or other that they identify elli with Virgil; most of them do so by omitting this pronoun in their paraphrase or translation of line 79: “se posuisse pedes ubi habebat caput ...” (Pietro di Dante) or “si capovolge; porta quindi la testa dove prima teneva i piedi” (Steiner; cf. also Benvenuto da Imola, Landino, Giovanni da Serravalle, Andreoli, delli Bargigi, Francesia, Venturi). Equally revelatory is the paraphrase of Porena, with its change of construction: “Virgilio ha volto la testa verso la direzione che prima era delle gambe ...” Scartazzini, instead of offering his own version, quotes the translation of Pietro di Dante; and Pfleiderer's (rather vague) comments accompany the translation of Streckfuss: “Wo erst der Fuß war, kam das Haupt zu stehn.”

The majority of the commentators who discuss line 79 are content to state merely that Virgil reversed his position on Lucifer's body (si capovolse, si rivoltò)—an interpretation that would hold true with either of the two possible references of elli (whether Virgil turned [started to turn] his head in the direction of his own legs, or of Lucifer's legs, he would be reversing his position). Given, however, the unanimous opinion of the translators (see above) in favor of the first assumption, it is highly likely that any commentator who does not take the stand of identifying elli with Lucifer is siding with the translators and with the majority of the commentators who have committed themselves. (Like Steiner, they may be reasoning, influenced by their own interpretation): “si capovolge; porta quindi la testa dove prima teneva i piedi.”

² Compare also the English translations of Anderson, Ayers, Butler, Cummins, How, Johnson, Langdon, Lockett, Lowe, Musgrave, Parsons, Plumptre, Shaw, Sinclair, Swiggett, Tozer, White, Wright.

³ Compare also the French translation of Masseron, and the anonymous 17th-century French translation.

⁴ Compare also the German translations of Bartsch, Kannegieser, Notter, Philalethes (pseud.), Ramhorlt, Streckfuss, Witte.

⁵ Compare also the Dutch translation of v. Velzen.

⁶ Though Meyer-Lübke translates Old It. zanca by 'Fuß,' the word is not found in this sense in Tommaseo-Bellini or Petrocchi. Corominas in his dictionary attests this meaning in Romance (e.g., Port. chanca 'pe grande') but not for Italian: in an anatomical reference the word means only 'leg,' just as is true of modern cianca.

⁷ There is a vast difference between 'where he had his feet (at the moment)' and 'where he had had his feet (earlier)'; when Dante has said the first, one hardly has the right to translate as if he had said the second. The Divina Commedia is not the Cantar de Mio Cid, whose tense system still baffles scholars.

⁸ The expression ‘anaphoric pronoun’ will be used throughout to designate a (subject) pronoun that repeats the subject of the immediately preceding clause.

⁹ Note also the rhyme: anche—zanche.

¹⁰ Originally this study was limited to the pronoun egli, since it is the tonic form that we find in our passage. Further

investigation revealed, however, that in terms of the categories set up for egli, there was no difference in the behavior of the two pronouns. Accordingly, they have been taken together; and the label egli will be used for both, without distinction.

¹¹ Not included among the examples of anaphoric egli are those in which the pronoun (a) is combined with a second subject:

che tutta ingrata . . .

si farà contro a te; ma poco appresso

ella, non tu, n' aurà rossa la tempia.

(Par. xvii.66)

or (b) is reinforced by means of stesso:

Quando si parte 1' anima feroce

dal corpo ond' ella stessa s' è disvelta . . .

(Inf. xiii.94–95)

It should be clear why, in this analysis of anaphoric pronouns for the purpose of comparison with our passage, only those of the third person have been considered. That the use of first- and second-person pronouns must differ from that of third-person pronouns is due to two factors (that work in opposition): on the one hand, the pronouns of the first and second person are rarely needed for identification, as egli rather often is; on the other, their highly personal reference lends itself to redundant use in emotional utterances, to an extent not matched by the more impersonal egli.

Incidentally it is interesting to note that, of the 103 examples of the third-person pronouns used anaphorically, only 30 represent the atonic form: the tonic egli is used more than twice as often as ei in an anaphoric function. This is not what one would be led to expect from Rohlf's treatment (ii.157–217), which gives rather the impression that it is only the atonic form that is used anaphorically.

¹² The phrase “immediately preceding predication” is meant to include not only dependent clauses, but also participles and infinitives that serve a similar function. Two limitations, however, have been introduced:

To be considered an “immediately preceding predication” it is necessary that (a) all the basic elements of the clause in question precede the one with egli; and (b) the two clauses be on the same level of discourse (either narrative or conversation). For the first reason, the following example is excluded: the antecedent of elli (line 20) is the subject of a clause that is interrupted by the elli-clause:

“Madonna, quelli che mi manda a vui,

quando vi piaccia, vole

sed elli ha scusa, che la m' intendiate.

(V.N. xii.18–20)

For the second reason, certain examples in which narrative and conversation are combined have been included even though the pronoun does not continue the subject of the clause that actually precedes it on the printed page: 'Questo superbo volle essere sperto di sua potenza contra al sommo Giove,' disse 'l mio duca, 'ond' elli ha cotal merto.

(Inf. xxxi.91–93)

. . . uno spirito d' amore . . . pingea fuori li

deboletti spiriti del viso, e dicea loro:

“Andate-voi a onorare la donna vostra”; ed elli rimanea nel luogo loro.

(V.N. xi)

Obviously a pronoun in conversation cannot have as antecedent a word of the narrative (Inf. xxxi.93), or vice-versa (V.N. xi).

It is not impossible, however, that the insertion of the alien clause, particularly when conversation is interpolated into narrative, may encourage the use of anaphora.

¹³ In other words, no example with anaphoric egli will be considered similar to our own unless its antecedent is non-generic and it is found in a clause that is dependent, is introduced by an adverbial conjunction, modifies an independent clause by modifying the verb of that clause (i.e., not an adverbial element, or a noun, e.g., la parte dov' ei son . . .), and follows the main clause that it modifies, instead of preceding it or being inserted into it.

That the features just listed are, in a general way, significant for syntax, should be immediately obvious; just why this particular combination of features should be decisive for the exclusion of an anaphoric pronoun is less obvious; a few explanatory suggestions will be offered later (note 21).

¹⁴ Cf. also Par. xxx.146; Purg. viii.115; ix.91; xi.69; xxvii.106; Inf. x.68; xxi.119; xxxi.93; xxxi.128; V.N. xix [p. 35 in Barbi's ed.]; xi [p. 40].—Incidentally, if Par. iii.85 quoted above is read ‘E’ n la sua volontate è nostra pace... ', then the ella of the next line would not be anaphoric.

¹⁵ Cf. also Purg. vii.50; viii.10; xxiv.37; Par. v.131; xxvii. 103; xxxi.113; Inf. iii.67; v.58; x.49; xi.74; xviii.88; xxi.21; xxi.135; xxiii.61; xxv.16; xxxii.115; V.N. xi [p. 41]; xxix [p. 123]; xxix [p. 124].

It should be stated that clauses introduced by the coordinating onde (cf. . . . ond' ei si maraviglia of Purg. vii. 11, quote above) will be considered independent, as will also those introduced by the “corroborative” chè, e.g.,

... e non pur a me danno

Superbia fa, chè tutti i miei consorti

Ha ella tratti seco nel malanno (Purg. xi.67–69)

¹⁶ Cf. also Inf. iii.90; vi.39; xii.81; xii.131; xiv.58; xviii.12; xxiii.18; xxiv.117; xv.30; xvii.12; xxix.105; xxxiii.3; xxxiv.34; xxiv.131; xxxiv.132; Purg. iv.91; ix.126; xi.44; xvi.101; xviii.127; xxiv.86; xxiv.144; xxviii. 126; xxix.138; xxx.45; xxxi.84; xxxii.49; Par. vi.73; ix.23; xii.19; xiii.122; xiv.15; xxii.57; xxiii.126; xxiv.36; xxv.65; xxviii.17; xxx.12; xxxi.33; xxxiii.135; V.N. xxi [p. 85]; xxiii [p. 96]; xxvi [p. 117]; xxvi [p. 119]; xxxvi [p. 144].

¹⁷ Cf. also Par. v.130–132; xxiii.124–126; xxxi.16–18; Inf. xi.76–77; xxvi.47–48.

¹⁸ Cf. also Inf. iv.34–35; xxi.64–66; xxxi.49–54; Purg. iii.53–59; iv.88–98; viii.91–95; xiv.49–51.

In the example quoted above “Quand' elli ebbe . . . compiuto,

/ La fiamma ... si partìo“ (and in every other example in this group except Purg. xiv.49–50) we have to do not only with the reverse of the order of clauses found in ”Volse la testa ov' egli avea le zanche,“ but with something else even more important: the (anaphoric pronoun) subject of the dependent clause differs from that of the main clause that it modifies. And in such cases (perhaps to avoid theoretical ambiguity, cf. Quand' ebbe . . . compiuto, / La fiamma . . . si partìo) the presence of the pronoun would seem to be the rule rather than the exception (at least when the verbs of the two clauses are in the same number and person): a study of third-person zero-subject in the Inferno reveals only one instance of the omission of the pronoun in the construction in question:

Lo mio maestro disse: “Questi è Caco ...”

Mentre che sì parlava, ed ei trascorse

E tre spiriti venner sotto noi (Inf. xxv.25–35)

Here, the pronoun was probably omitted from the mentre che clause in order that the ei of “ed ei trascorse” (indicating the shift of subject to Caco) might stand out in greater relief.

¹⁹ Cf. also Par. vii.115–117; xi.55–57; Inf. xxiii.52–54. That anaphoric egli may appear in a clause that modifies an adverbial element of the main predication may be explained by the fact that, in such cases, we always have to do with a correlative construction (si . . . che, più . . . che etc.), i.e., a

parallelistic construction (which seems to favor anaphora: see note 21). And if it tends to be excluded from a clause that modifies the verb of the main predication, this is perhaps because such a clause is in closer connection with the subject of the main clause (since the subject governs the verb, not an adverbial element): that is, in closer connection with the antecedent of its own subject. In “...” disse Minos a me quando mi vide, the main verb disse, governed by Minos and determined by quando mi vide, serves, as it were, to transmit the essence of Minos directly to the verb vide of this determining clause—which thus has less need of an explicit subject of its own.

²⁰ There is a second way (one, incidentally, much easier to demonstrate) of proving the non-anaphoric reference of elli in Inf. xxxiv.79, and that is by reference to word-material instead of construction: to the individual word ove that introduces the clause containing the pronoun. The frequency of anaphora (found mainly in dependent clauses) varies according to the conjunction used: the greatest number of cases are found with se, come, and che+preceding correlative, while none has appeared with poscia che, da che, sì 'until.' As concerns ove (and the same is true of dove) it may be said that whenever the pronoun appears in a clause introduced by this conjunction, IT NEVER REFERS TO THE SUBJECT OF THE MAIN CLAUSE. This is, of course, necessarily true when the overall construction is that of Lo duca . . . le zanche (e.g., “Questo m' avvene ovunque ella mi vede”: V.N. xxvii); it happens also to be true in every other case, regardless of construction. Whether (d)ove introduces a noun clause:

Questa ballata in tre parti si divide: ne

la prima dico a lei ov' ella vada . . . (V.V.xii)

or an adjective clause:

... la parte dov' ei son rende figura

(Inf. xvii.12)

or an adverbial clause which modifies not the verb but the adverbial element là (and follows a dependent rather than an independent clause):

Ei mormorava, e non so che ‘Gentucca‘

sentiva io là ov' ei sentia la piaga

della giustizia che sì li pilucca.

(Purg. xxiv.37–39)

the pronoun always refers to an entity different from the subject of the main clause (Cf. also Inf. xxiii.6; xviii.73; Purg. iii.132; xxxii.108; Par. xi.49; V.N. xxi [p. 85].)

And by the same token, when the subject of the two clauses is the same, the pronoun is never found—again, regardless of construction:

Con quel furore e con quella tempesta

ch' escono i cani addosso al poverello,

che di subito chiede ove s' arresta

(Inf. xxi.67–69)

... e però nel secondo

giron convien che sanza pro si penta

qualunque priva sè del vostro mondo . . .

e piange là dov' esser de' giocondo

(Inf. xi.41–45)

Quante il villan . . .

vede lucciole giù per la vallea,

forse colà dove vendemmia ed ara . . .

(Inf. xvi.25–30)

(Cf. also Inf. xv.114; xx.78; xxv.72; xxv.94; xxvii.48 Purg. i.23; ix.36; xiv.34; xviii.30; xvi.15; xviii.102; Par'

vi.72; xxii.66; xxv.95; V.N. xv [p. 62]; xxi [p. 86]; xli [p. 162].)

This does not necessarily mean that anaphoric egli is excluded with (d)ove; in the following example the pronoun repeats the ella of the preceding (dependent) clause:

Ne li occhi porta la mia donna Amore,

per che si fa gentil ciò ch' ella mira;

ov' ella passa, ogn' om ver lei si gira

(V. N. xxi.1–3)

It does not, however, refer to the subject of the clause modified, which is ogn' om (this shift of subject regularly calling for presence of the pronoun in the preceding clause, see note 18).

Thus two definitive statements may be made that exclude the possibility of lo duca serving as antecedent of elli in Inf. xxxiv.79: (a) “in the construction represented by lo duca ... le zanche the third-person pronoun is never used anaphorically”; and (b) “in a clause introduced by (d)ove the third-person pronoun never refers to the subject of the main clause.”—When a grammatical analysis is correct it can often be proved by reference to more than one criterion.

²¹ In this treatment of pronominal usage in the Commedia (and Vita Nuova), linguistic analysis has played an ancillary role: that of demonstrating the impossibility of a certain translation of a certain line of poetry. Any linguist must accept the conclusions here reached (unless he can prove faulty analysis, and crucial faulty analysis, of the material classified), but no linguist should be satisfied with the present treatment as an adequate description of pronominal usage. It is true that we know two reasons why anaphoric egli is excluded from Inf. xxxiv.79, but we do not know why these reasons should be effective, and we have no idea what other reasons may exist for exclusion, in general, of anaphoric third-person pronouns; again, we know, in terms of a mainly unsifted classification, when the presence of anaphoric egli is possible, but we have no idea whatsoever when it is necessary. A far more detailed analysis of the evidence presented, as well as a detailed analysis of new evidence (verbs with zero-subject), would be necessary before we could hope to begin to understand the phenomenon which is here treated as a means to an end. And such an involved linguistic study would be a project in itself, out of place in this article devoted to the interpretation of Inf. xxxiv.79.

It may not, however, be out of place to point out, by way of rounding off our discussion of the linguistic problem represented in this line, three widespread tendencies characterizing the material collected:

(a) In 25 out of 66 cases the subject of the dependent clause which serves as antecedent to egli is a relative pronoun; it may refer to a specific entity (“Questo diss' io diritto alla lumera / che pria m' avea parlato, ond' ella fessi / lucente . . .”: Par. v. 130–132) or to a type (“ ... chi volesse / salir di notte, fora elli impedito / d' altrui?”: Purg. vii.49–51). It is surely true that a relative pronoun distinguishes itself from other possible antecedents in that it can never be repeated in the following clause: “John goes where he wants” can conceivably be paraphrased by repetition of the antecedent (“John goes where John wants”), but it would be utterly impossible to imagine the same kind of paraphrase for “This boy who goes where he wants ...” (“This boy who goes where who wants ...”). It may be because of the abstract quality of such an antecedent that the pronoun is so often felt necessary, by compensation, in the following clause. (And the predilection for anaphora following a generic antecedent,

even when not a relative pronoun, may perhaps be similarly explained.)

(b) Of the 71 cases in which egli is found in a dependent clause, in 27 this is a relative clause with oblique che (quanto). Frequently we find a combination of (a) and (b); this is what we have in the following example, with two instances of (b): “... d' un ruscelletto, che quivi discende / per la buca d' un sasso ch' elli ha roso / col corso ch' elli avvolge . . . ”: Inf. xxxiv.130–132. At first thought it might be surprising that the relative pronoun would tend to attract anaphoric egli, when (d)ove, which is a relative conjunction, reveals the opposite tendency. But if one remembers that the form che (quanto) is ambiguous, serving either as subject or as object, one instantly comprehends that the presence of the subject egli serves to point conclusively to the oblique function of che.

(c) Apart from these two constructions we find, in all four groups illustrated above in the text, the tendency to use egli when a parallelism of some sort is intended. Occasionally, there is a comparison between the acts (states) of two different entities (“Ma poco poi sarà da Dio sofferto / nel santo offzio; ch' el sarà detruso / là dove Simon mago è per suo merto”: Par. xxx.145–147: cf. also the type with double subject, note 11); more often, however, we find a parallel between two acts (states) of the same person:

‘S’ ei fur cacciati, ei tornar d' ogni parte,'

(Inf. x.49)

S' el fu sì bello com' elli è or brutto .. .

(Inf. xxxiv.34)

'Se la vostra memoria non s'imboli

nel primo mondo da 1' umane menti,

ma s' ella viva sotto molti soli,

ditemi chi voi siete . . . (Inf. xxix.103–106)

Io sono Omberto; e non pur a me danno

superbia fe', chè tutt' i miei consorti

ha ella tratti seco nel malanno.

(Purg. xi.67–69)

Io dico che, secondo 1‘ usanza d’ Arabia,

1' anima sua nobilissima si partiò ne la prima

ora del nono giorno del mese; e secondo 1'

usanza di Siria, ella si partìo nel nono mese

de l' anno . . . (V.N. xxix)

This is surely the same tendency that is reflected in the frequency of anaphoric egli in correlative constructions; in either case, the presence of the pronoun serves to maintain a balance: since the two predications are of equal weight, each should have its subject expressed.

²² Why does Dante not save his diatribe for the sinners in the ninth and last bolgia, whose crimes have been classified by Virgil, his teacher, as the most heinous? Enrico Sannia (Il comico . . . nella Divina Commedia, p. 162) suggests: “Se la frode occupa il più basso loco nella sua morale teologica, la cupidigia del clero ha il posto più abbietto nella sua morale sociale.”

²³ Cf. the forthcoming article, “The Aesthetic Structure of Inferno xix,” by Mark Musa.

²⁴ Thus zanche with Dante is restricted to legs that are raised upright, emerging from the ground, unable to perform their natural function of movement through space. It is true that in Inf. xxxiv.79, where the word is applied to Lucifer, his legs have not yet been presented to us in this position; but the next moment they will be, they must be; and to Virgil, when he turned his head, it was thus that they appeared. His pupil, through whose eyes we must see, was slow to orientate himself.

In choosing the word zanca, the poet may not only have been influenced by its pejorative connotations: he may have been familiar with the variant meaning of zanca ‘asta’; and it is surely in an upraised position that legs, particularly naked legs (as we must assume to be those of our sinners), must closely resemble sticks (cf. xrx. 47: “... come pal commessa”). Indeed Machiavelli, in his charming dialogue with Dante which he inserts into his Discorso o dialogo intorno alla nostra lingua, has Dante explain as follows his choice of zanca in Canto xrx:

N. Dimmi: tu di' ancora, volendo dire le gambe,

E quello che piangeva con le zanche,

perchè lo di' tu?

D. Perchè in Firenze si chiamono zanche quelle aste,

sopra le quali vanno gli spiritelli per san

Giovanni, e perchè allora, e' 1' usano per gambe;

e io, volendo significare gambe, dissi zanche.

As a matter of fact, Dante would seem to be saying that for him the word zanca had only the meaning ‘asta’ and that he personally was the creator of a semantic shift: ‘asta’ >‘gamba.‘ Of course, this must be taken in the spirit of Machiavelli's dialogue.

²⁵ The first impression that comes to the reader of Canto xxxiv as he senses the parallel in question must be that of “post-figuration” rather than prefiguration: in reading Canto xix (for the first time) we obviously cannot see Nicholas as prefiguring Lucifer; only in Canto xxxiv can the parallel between the two be sensed, and the first reaction (seeing Lucifer and remembering Nicholas) must be: “Lucifer is in the same position as Nicholas!” It is only when rereading, or re-thinking, Canto xix (seeing Nicholas and remembering Lucifer) that we can truly see the simonist as the prefiguration of Satan, see his figure against the background of the Devil's colossal shadow—which is surely the way Dante would have us see him: it is he, alone, who gains in significance from the parallel.

²⁶ It is true that there are an indefinite number of upraised legs in the same circle with Nicholas, but he is the only one who is individualized and the only one seen at close range by the pilgrim.

²⁷ In addition to the direct correspondence between Nicholas' contrappasso and that of Lucifer, who are both buried in the ground upside down, Canto xxxiv also contains several variations on this contrappasso. The legs (and torso) of Judas emerge from the chasm of Lucifer's central mouth, while Brutus and Cassius both hang head downward from the mouths on both sides; as for the anonymous sinners, some are standing on their heads and others, bent like a bow, have their heads (and feet) on the ground. Moreover, in the description of Brutus, there is also a verbal parallel linking this

inhabitant of Lucifer's circle with Nicholas: lines 64–66 of our canto,

Degli altri duo ch' hanno il capo di sotto

Quei che pende dal nero ceffo è Bruto—

Vedi come si storce, e non fa motto

are surely intended to recall the lines (46–48) of Canto xrx in which the pilgrim addresses Nicholas:

‘O qual che se’, che 'l di su tien di sotto,

Anima triste, come pal commessa,'

Comincia' io a dir, ‘se puoi, fa’ motto.'

Of these correspondences only the first has apparently been seen by the commentators: according to Longfellow (i.437) the similarity between the position of Judas and that of Nicholas was pointed out by D. G. Rossetti in his Spirito antipapale (i.75, Miss Ward's tr.); of the later commentators consulted, only Grandgent has seen fit to comment on the parallel.

²⁸ Thus it is difficult to understand why this parallel has escaped the attention of all the commentators up to the present.