manuscripts of the tradition. Moreover, he draws attention to the contexts into
which the third and the fourth interpolations are interwoven and concludes that
these passages of foreign text are remarkably more fitting in the narrative con-
texts of A, B, and C, rather than in E.120 His stemma is shown in Figure 3.
The Manuscript Filiation of Niðrstigningar saga 51
Latinsk tekst
norrønt arketyp
a1
b
C
a2
E
B
a3
A
D
Figure 3. Odd Einar Haugen’s stemma
The Present Stemma
As shown in Figure 4, A, B, C, D, and E share a common archetype α, which
includes the four interpolations typical of Niðrstigningar saga. From α, a lost
manuscript β was copied, and from this C is ultimately derived. C stands alone
in a distinct branch of the stemma, separated from the rest of the tradition, and
in view of its absence of errors must be considered a more reliable and stable
text. C also seems to be the closest of all manuscripts to α. As a matter of fact,
rather than the textual corruptions shared by A, C, and D, the three possible
conjunctive errors – “oc for þangat” (“and [he] travelled there”), “slita ondina”
(“[to] tear away the soul”), and “maþr allosęligr” (“a most joyless man”) –
seem more likely to be secondary innovations exclusive to B. They may simply
have been readings already present in α, from which they were transmitted to
52 Niðrstigningar saga
Latin T
O
α
γ1
β
γ
B
Latin K (Majority Text)
C
E
δ
A
D
Figure 4. The present stemma
the two lost exemplars, β and γ, and from there to the daughter manuscripts.
This evidence excludes Magnús Már Lárusson’s hypothesis that held B to be
the closest variant text to the archetype of Niðrstigningar saga.121 The copy γ
introduced two new errors – “þa er ec lifða” (“when I lived”) and “sa hafði”
(“who had”) – which separates it from β. The errors typical of γ were then
transmitted to one of its lost copies, δ, which along with another error intro-
duced by δ alone – “scolom biþa” (“we shall wait”) – was transmitted to the
two sister manuscripts A and D.
This picture is then complicated by the vicissitudes of B and E, which are
also derived from γ but, on the basis of the evidence discussed, seem to have
undergone secondary revisions. B embellished some of the original readings of
γ (possibly ex ingenio), harmonizing them to its taste and possibly conforming
and adopting them to their literary context. This revision of γ is indicated
The Manuscript Filiation of Niðrstigningar saga 53
below as γ1. As shown in the collations above, it seems clear that text E also
descends from γ on account of the vicinity of its readings to those of A and B
within the interpolated portions of text. However, there are indications that the
text of E was subsequently recorrected (this time ex libro) on the basis of an-
other exemplar of the Latin Evangelium Nicodemi, which did not pertain to
Latin T. The text of E seems instead to have been corrected on the basis of a
codex displaying features of the Majority Text K, as deduced from their agree-
ment of readings below.122 It is therefore appropriate to refer to an older redac-
tion of Niðrstigningar saga, represented by A, B, C and D, and a more recent
fully-revised redaction of the text, represented by E alone.
The two possible conjunctive errors of A and B and A and C cannot be con-
clusive evidence of their close relationships, as the readings in question are
parts of common canticles and psalms abundantly sung in a monastic environ-
ment and could therefore have been easily emended during the time of the
transcription of C and D. From the evidence discussed, it is clear that C is the
closest surviving codex to the archetype, which is evidently contaminated due
to the secondary revisions of B and E, ex ingenio and ex libro respectively.
Accordingly, the new stemma can be implemented as shown in Figure 4.
4 The Latin Source Text Underlying
Niðrstigningar saga
The text of Niðrstigningar saga has hitherto been related and compared to the
so-called A version of the apocryphon, representing the Majority Text (K) of
the Latin tradition.1 This version was by far most widely diffused in medieval
Europe as, judging from their incipit, over 380 out of the 434 counted wit-
nesses transmit the Majority Text either fully or in an abridged form.2
Despite the overall agreement of readings, lexicon, and style, the Majority
Text does not fully represent the Old Icelandic rendition. Already after a first
collation, it seemed clear that the older redaction of Niðrstigningar saga exhib-
ited major and minor narrative details typical of the hybrid redaction (T) rather
than the more conventional readings of K. Evidence of their dependence is
confirmed by several textual and thematic correspondences, as illustrated by
the collations of readings below. In addition to T and K, the collation also com-
prises the readings of the R text, the sole surviving Icelandic exemplar of the
Latin Evangelium Nicodemi, Reykjavík, Þjóðminjasafn Íslands, 921, a frag-
ment dating to the thirteenth century that also transmits readings typical of the
Majority Text.
The Icelandic text is represented in the collations by the readings of A (AM
645 4to), which is the oldest surviving manuscript to transmit the older redac-
tion of Niðrstigningar saga and the only medieval codex to preserve the text in
its entirety. Its readings are collated and tested with those transmitted by the
younger redaction E (AM 238 V fol.), a single fragment leaf copied during the
fifteenth century. Its text ultimately derives from the same archetype as recen-
sion A but with subsequent corrections and revisions based on a Latin exemplar
transmitting the Majority Text.3 Aside from its distinctive textual divergences
from the Latin source text – such as evident omissions, abridgements, and rear-
rangements of the plot – the older redaction reflects, to different extents, both
major and minor textual features typical of T.
The Latin Source Text Underlying Niðrstigningar saga 55
In contrast with the Majority Text type, the hybrid redaction rephrases and
reformulates the original text several times and concurrently substantiates the
plot with dramatic details and anecdotes. These new additions, all varyingly
mirrored in the Icelandic translation, must have originated within the hybrid
text as important logical threads, apt to integrate and develop the narrative of
the original pseudo gospel when it was perceived as either limited or deficient.
During the twelfth century, secondary revisers might have intervened to adjust
and amplify certain passages of the Majority Text, enriching it with details of
Christ’s Descent and Harrowing of Hell derived from other popular religious
/> narratives concerning the same catabasic theme (such as homilies and ser-
mons) and also possibly adding realistic details from visual art pieces that de-
pict this prominent scene.4
In general terms, the Majority Text is characterized by constant stateliness
and sobriety in the treatment of the Christ figure and seems to avoid any de-
tailed description of his Harrowing of Hell or of the military imagery tradition-
ally associated with it. The hybrid redaction seems to compensate for this
absence of action by adding more traditional iconographic details to the figure
of Christ, hence giving a much more vivid and dramatic force to his arrival in
Hell and his dealings with Satan.
Although highly elaborated and adapted to a new literary context, five major
textual digressions of Niðrstigningar saga (confined in a short emphatic pas-
sage corresponding to the end of chapter XX1.3 and the beginning of chapter
XXII.1), absent from the Latin Majority Text, find thematic and formal corre-
spondences in T. In order of their appearance in the text, they involve Christ’s
physical shattering of the gates of Hell; a description of a host of angels attend-
ing him; his breaking of the infernal bonds that chained the patriarchs and
prophets in Hell; the astonishment of the inhabitants of Hell at his sight in their
realms; and his physical binding of Satan. As will be shown below, these de-
tails are either absent or highly elusive in the Majority Text. The other textual
differences between the Majority Text and the hybrid redaction are minor and
mostly concern a different choice of lexicon or alternative wording.
The Prologue
The prologue of the Majority Text lacks the assertion typical of T, which alleg-
edly ascribes the commission of the first translation of the Hebrew apocryphon
to the Emperor Theodosius the Great (†395). Its prologue simply ends by at-
tributing the authorship of the text to Nicodemus: K “Mandauit ipse Nichodemus
litteris ebraicis” (“This same Nicodemus wrote it himself in the Hebrew script”).
56 Niðrstigningar saga
Figure 5. Illumination attributed to the Master of the Parement of Narbonne in
the Très belle Heures de Notre-Dame (Paris, BnF, nuov. acq. lat. 3093), f. 155r,
lower side (ca. 1375–1400) depicting the Harrowing of Hell as related in the
Gospel of Nicodemus. Reproduced with permission of the BnF.
The Latin Source Text Underlying Niðrstigningar saga 57
K Prologue 13/13–14
T 90r/9–11
A 54v/20–3 (Epilogue)
Mandavit ipse Nichodemus
Ipse Nichodemus scripsit in
Enn morgom mannzøll⟨drom⟩
litteris ebraicis.5
litteris hebraicis. Theodosius
siþar comsc at boc þeire
autem magnus imperator fecit Theodosius keisere sonr
ea transferri de hebreo in
Archadii. Hann hafði meþ
latinum.6
sér i Miclagarþ oc let þar
uppraþa oc varþ þar monnom
alldat umb.7
The additional detail found in the prologue of the hybrid redaction could
have originated as a secondary development of the prefatory rubric shared by
the entire Latin tradition. Due to the swift assertion that the text was discovered
in Jerusalem during the reign of Theodosius the Great, the compiler of the hybrid
redaction must have made the consequent assumption that its translation into
Latin was made at the personal instance of the Emperor: K Rubric 13/1–4
“IN NOMINE SANCTAE TRINITATIS INCIPIUNT GESTA SALUATORIS
DOMINI NOSTRI IESU CHRISTI INUENTA THEODOSIO MAGNO
IMPERATORE IN HIERUSALEM IN PRETORIO PONTII PILATI IN
CODICIBUS PUBLICIS.”8
Notwithstanding the unambiguous reference to the Emperor Theodosius I as
the owner of the text and commissioner of some literary initiative to promote
the circulation of the apocryphon – which clearly establishes a close relation-
ship of the Icelandic text to T against the Majority Text – it is worth noting
the unusual choice of the verb “uppráða” (“to read aloud”), which renders the
Latin verb “transfero” (“translating, interpreting”) of T.9 The Icelandic com-
piler removes this information from the prologue and briefly recalls it in the
epilogue of the narrative, which follows a considerably abridged version of
chapter XXVII.5, a section describing how, after hearing the dramatic ac-
counts of Carinus and Leucius, Joseph and Nicodemus reported to Pilate
their accounts of Christ’s Descent into Hell.
K XXVII.5 49/1–5
T 103v/27–8
A 55v/19–20
Haec omnia quae dicta
Hec omnia que dicta sunt et
Nicodemo oc Iosep⟨h⟩
et facta sunt a Iudaeis in
facta a Iudeis in synagoga
enn þeir reþo upp ⟨fyrer⟩
sinagoga eorum statim
eorum Ioseph et Nichodemus òðrom´.12
Ioseph et Nichodemus
anuntiauerunt presidi
adnunitiauerunt presidi.10
Pylato.11
The semantic correspondence of the Icelandic verb “uppráða” (“to read
aloud”) with the verb “annuntiare” (“to announce”) is considerably stronger in
this context than in the previous case, as the Latin verb “annuntiare” also im-
plies a public speech or exposition (“to reveal / to proclaim / to declare pub-
licly”). The use of the verb “uppráða” in this passage, immediately preceding
58 Niðrstigningar saga
the reference to Theodosius, may have been at least partially influenced by its
second employment in the following sentence. The Emperor’s request to trans-
late the Hebrew pseudo gospel is adopted into Joseph and Nicodemus’s report
of the previous sentence and is converted into a request for a simple public
reading of the text. A second, more reasonable explanation of the divergence of
the readings T “transferre” and A “uppráða” can also be seen as a deliberate
attempt of the Icelandic compiler to conform the epilogue of Niðrstigningar
saga to that of the Jarteinabók Þorláks byskups in forna (“The Ancient Miracle
Collection of Bishop Þorlákr ”). In the latter, the first and foremost item of AM
645 4to (ff. 1r–11v), the unusual verb “uppráða” – rather than the considerably
more common synonym “upplesa” – is also employed for the public reading of
Þorlákr’s miracles at the Althing of 1199.13
The Shattering of the Gates of Hell
The Majority Text remains silent about Christ’s physical shattering of the gates
of Hell and mentions only briefly that Christ “broke the indestructible bonds”
(K XXI.3 41/18–19 “insoluta uincula disrupit”) of the souls of the righteous
with the aid of his “unconquered power” (K XXI.3 41/19 “inuictae uirtutis”).
The beginning of chapter XXII.1 is emblematic of this discretion and solici-
tousness in the treatment of the figure of Christ; here, it is clear that only at the
sight of Christ in Hell do its inhabitants hurry to fully acknowledge his victory:
K XXII.1 42/1–5 “Haec uidentes Inferus et Mors et impia officia eorum cum
crudelibus ministris expauerunt in propriis regnis agnitam tanti luminis
clarita-
tem dum Christum repente in suis sedibus uiderunt, et exclamauerunt dicentes:
‘Uicti sumus a te.’”14 This brief description of Christ’s effortless victory over
the inhabitants of Hell does not seem to have convinced or amused the compiler
of the hybrid text. He expanded and magnified this image in the following sec-
tion, XXI.3, creating a much vivider and fiercer scenario that describes Christ’s
physical destruction of the gates of Hell in detail. At Christ’s appearance in
Hell, “all the infernal gates, bars, and locks were destroyed” (“omnes porte in-
fernales et uectes et sere in accessu eius confracte sunt”), and “everything gave
Him space and made room for Him” (“omnia locum illi et uiam dederunt”).
This additional scene within the hybrid text concerning Christ’s physical de-
struction of the gates of Hell was most probably influenced by the wording of
Psalm 107(106):15–16, transmitted in chapter XXI.2. This section immediately
precedes Christ’s destruction of the gates of Hell and has King David recall in
his own words one of the psalms commemorating the power of the Redeemer:
“Let them confess to the Lord His mercy and His wonders to the sons of men,
The Latin Source Text Underlying Niðrstigningar saga 59
because He has shattered the gates of bronze and destroyed the bars of irons”
(T 101r/8–10 “Confitaentur domino misericordie eius et mirabilia eius filius
hominum quia contriuit portas ereas et uectes ferreos con⟨f⟩regit”).
The Icelandic translator further emphasizes Christ’s warfaring spirit and
physical strength, especially in his peculiar choice of vocabulary describing the
fortress-like architecture of Hell. Christ, it is said, “arrived to the stronghold of
Hell” (“com at helvitis virki”); he abruptly “destroyed” (“braut”) “the fortress
of Hell” (“borg helvitis”) and finally opened up “a large gate” (“hliþ miket”).
The choice of the word “fortress” (“borg”) in this passage prefigures the de-
scription of the innermost infernal fortified compound, the “inner fortress of
the prison” (T 102r/4 “ima carceris claustra”) of chapter XXIII.1, which Christ
is doubtless able to destroy.15
K XXI.3 41/16–20
T 101r/26–9
A 54r/13–16
Superuenit Rex Gloriae in
Filius Dei Christus Rex
Niorstigningar Saga Page 10