Younger Runes in Manuscripts and Early Printed Works

Stowe MS 57, British Library

This compilation manuscript is tentatively ascribed to Geoffrey of Ufford who flourished around 1154. Close to the beginning of the manuscript, a collection of scripts is given: Hebrew and Greek on fol. 3 recto, Latin and younger runes on fol. 3 verso and the alphabet of Aethicus Ister on fol. 4 recto (the recto pages are numbered ıııı· and v· respectively, but there is now no leaf between ı· and ııı·). High resolution images of the manuscript are available at the British Library manuscript viewer.

Runic material

The runic material occupies the lower part of folio 3 verso. It follows a Latin alphabet with letter names, which lists & et and ÷ titel as well as the originally Greek additions Y and Z as part of the normal alphabetical sequence. These are immediately followed by the subheading Anglıcę lıꞇꞇ̃ę, normalised Anglicae litterae English letters; and then the letters Ƿ, Ð, Þ and ⁊. Style and layout confirm that these form a part of the original content, whereas the entry between titel and the subheading is a later addition, possibly meant to illustrate the use of titel.

Spanning the entire runic section is the heading Lıꞇꞇerę|normoꝛv̅|qvę ꝺıc|vn|ꞇur rvɴ ſꞇa-|feſ; normalised Litterae Norm[ann]orum quae dicuntur runstafes Letters of the Normans which are called runestaves. There is no abbreviation sign indicating the missing letters of Normannorum, so they must simply be inferred from context. If the heading was copied from a written source, this might have had a nasal bar over the m which was overlooked by the copier, but one would still expect the copier to have supplied the missing part in the same way as any reader would.

The content is placed in a table with seven rows and seven columns, but as there is no relation across the cells in the same row, this should be regarded as a linear sequence of entries only incidentally laid out in columns, and consisting of three sections written consecutively. They are indicated by subheadings added after the entries themselves, in the same ink as the main heading and presumably added by the original scribe as part of the original composition. They correctly correspond to the breaks between material of different origin, but do not provide additional information about their respective sections. Secondary material have been added to the first section in three phases, and the addition of a small fourth section seems to be contemporary with the latest of these.

As each of the three sections are in fuþark order, I have for clarity transcribed everything except the youngest phase of additions into a reorganised table with one row per fuþark entry. The non-linear contents of the first section is split into three colums for the original content plus two more for secondary additions. Also, positions are left blank where section 2 skips entries in the fuþark order. The resulting alignment between rows across the three sections is thus only implicitly present in the manuscript.

f	fe	f	3/1	f
u	oc ur	u	3/2
þ	oc þurſ		3/3
o	oſ	o	3/3+2
r	oc reıð		3/3+2
c	oc cogen		3/3+3
h	hachel	h	2/1
n	oc nouð	n	2/2
i	oc ıſ	ı	2/3
a	oc ar	a	2/3+1
s	oc ſol	ſ	2/3+2
ꞇ	ꞇẏr	ꞇ	2/1
b	oc bırken	b	1/2
m	maꞇher	m	1/2+1
	oc loꞇher	i	1/2+2
ẏ	oc ẏr	ẏ	1/3+2

Section 1, fuþark with sound values and names

The first column reproduces the runes of the first section, which are written in large size in the lower left corner of their cells. They are quite uncorrupted except for the adaptation to being written with an angled pen-nib. The forms are typical intermediate type runes, except for the o-rune which has the long-branch form, possibly erroneously as will be discussed below. The s-rune has the short-twig form with a dot at the bottom which is a not uncommon variant for this type.

Immediately following this are the values which are given in normalised form in the second column. The value is missing for the l-rune, but could be inferred from the rune name. The letters vary arbitrarity between normal lowercase letters and small capitals with no apparent meaning behind the distinction. Clear examples of small capitals are r, s, b and m, and I also interpret i and ẏ as such. For o and c it is impossible to tell the difference, and l is missing, which leaves f, u, þ, h, n, a and ꞇ as clear examples of lowercase letters. Five of the letters, u, r, c, s and m are followed by a full stop; also this seems to be entirely arbitrary, although disproportionally often occurring with small capitals. Using þ as the sound value of rune 3 and y of rune 16 reveals that the author is an Anglo-Saxon.

The value y for the sixteenth rune might seem to suggest that the ultimate source of the material is younger than the merger of /r/ and /ʀ/. However, the sound values in some other manuscripts are demonstrably derived from the rune names; this can therefore not be relied on as evidence, unlike the converse situation where the sound value is given as r. This means that it is not possible to tell whether the value o for rune 4 represents the younger usage transliterated o (covering /o/ and /ø/) or the older ą (covering /ã/, /æ̃/ and /ǫ̃/), as the latter also typically is represented by o in non-Norse manuscripts.

Above the rune and sound-value pairs, the rune names are given. As they are written in smaller letters I read them as subordinate in importance; they are therefore placed in the third column here despite strictly speaking occurring before the former two entries. The presentation of the names is quite peculiar, in that it has oc, Old Norse for and, in front of most names. Both this and a single use of the letter k in one of the names indicates a written Old Norse original source. Except for the first name, all names without a preceding oc is offset to the right to align with the other names. The pauses suggested by the lack of oc divides the list into five parts where the first two corresponds to the first ætt, the third group to the second and the last two to the third ætt. That the source understood the division into three ættir is otherwise obvious from the presence of cryptic runes in the same material.

Apart from orthographical details, only one name clearly seems to be corrupted: lother is unexpected and most easily explained as a dittography based on the preceding mather. The language is West Norse with preserved diphtongs in reið and nouð for normalised nauð or nauðr. The quality of the epenthetic vowel in mather for normalised maðr points towards a Scandinavian source rather than an Icelandic one. A parallel to writing ou for Old Norse au can be found in St. John’s College, Oxford, MS 17. The diphtong is lost both in cogen for Old Norse kaun and in the presumably corrupt name for the l-rune. There is no obvious explanation for the g in cogen, but the presumable epenthetic vowel suggests that a consonantal pronunciation lies behind it. The ch in hachel is easier to explain as an orthographic vagary. The only other name-form that is not straightforward is birken. A West Norse form of this late date would be expected to exhibit breaking, as in normalised bjarkan. The e in the second syllable contrasting with expected a makes it possible that this is a side form where the original short e in the first syllable was raised to i before breaking took place, or that the name had been reinterpreted as the related adjective birkinn made of birch which exhibits this development.

Secondary variants to section 1

Despite the smaller letter size, oc plus the rune names requires more horizontal space than a single rune plus its sound value. The resulting empty space in the lower right of the cells have in ten instances been filled with an alternate form of the rune following the sign ꝉ, the conventional abbreviation for Latin vel meaning or. In two instances a second alternative have been crammed in above these; and as it is obvious that the first ones were added without the expectation of a second one about to be added, the additions must have been made in at least two different phases.

In both cases where there are two variants, the one written first is the correct long-branch version of the rune, and the other a non-runic sign of uncertain origin. All other variants also fall cleanly into one of these categories, and I have placed them accordingly in the restructured table above. For the b-rune, there is no difference between the long-branch and intermediate types, but the variant here is merely drawn with more pointed loops. Apart from the two runes where two variants are given, there is little to confirm this division into two phases; but at least it looks as if the variant given for the i-rune is written after the second variant for the a-rune.

It is possible that the first phase was already present from the outset, but one would then expect e. g. ᛐ ꝉ ᛏ ꞇ rather than ᛐꞇ ꝉ ᛏ, and it would also fit less well in with the organisation of the surrounding material. On the other hand, if it was part of the original content, it might be an explanation for the long-branch form of the o-rune.

Section 2, Nemnivus’ alphabet rearranged in fuþark order

The second section is introduced by the subheading alıvꝺ, normalised aliud (an)other. Being singular neutral it cannot refer to the feminine plural litterae of the main heading, but must refer to an implied alphabetum or the abecedarium of the heading to the Latin alphabet above.

Its content is quite unique in all of the Runica Manuscripta material, in that takes the form neither of runes in fuþark order nor of runes rearranged in alphabetic order, but rather of an alphabet rearranged in fuþark order. Now, the alphabet in question is the invented alphabet of Nemnivus, Nemnius or Nennius, which as Derolez observes is in its visual appearance based on Anglo-Saxon runes and also primarily is preserved in close connection with this script. Nonetheless, it is otherwise always presented in alphabetical order, and with a selection of sound values based on the Latin alphabet. The names of the letters – which are not present here – are also more similar to the names of runes than to the ones of the Latin alphabet, and Derolez speculates that they might be derived from the Old English rune names.

Each of the thirteen cells contains a letter and a sound value, organised in fuþark order except skipping þ, r and k, and with a second i in the position of l. The omission of þ is of course due to the source lacking a letter for this sound, but that is not the case for the other two. In the table above, I have introduced gaps for the skipped letters so that the remaining ones line up with their equivalents in the other sections. In contrast to the first section, the sound values are here all in normal lowercase letters; except for the second i which looks more like a small capital, and as due to its position was undoubtedly ultimately meant as an l.

Derolez identifies two pairs of manuscript sources for this alphabet, and the members of each pair correspond closely enough to each other that he treats them as one in the presentation of the material. The first pair, which he refers to as N₁ consists of Oxford, Bodleian Library MS. Auct. F. 4. 32, where it is presented along with its origin story, and Cotton MS. Titus D 18 where it it immediately follows the Anglo-Saxon runic alphabet in a redaction of the De inventione text which is extended by among other thing a well-preserved younger runic alphabet with supplemental runes. The second pair, N₂, consists of the confused jumble that is St. John’s College, Oxford, MS 17, where it is interrupted by the best preserved of its Anglo-Saxon fuþarks, and in the lost Cotton MS Galba A 2 which is preserved with unknown accuracy in George Hickes’ Thesaurus, pars tertia.

The present version differs more from each of these than they do from each other, and is in some details more similar to the first and in others to the second. It must therefore derive from the hypothetical ancestor N₀ by a different route, deserving its own designation of N₃. This suggests at least one lost intermediate between this manuscript and N₀, yielding the following stemma:

*N₀
*N₁		*N₂		*N₃
Bodleian	Titus	St. John	*Galba	Stowe
			Hickes

Because of the complex relationship between the five, I find it necessary to present tracings of the characters, with normalised transcriptions of their sound values and, where present, their names. Due to the state of preservation of the two manuscripts in the N₁ group, the tracings of their characters fill in gaps where the ink has flaked off. Mostly, ink remains or discolourations reveal where lines have once been present, but in places I have let the internal logic of the strokes guide me if this corresponds to the sister manuscript. Likewise, some letter names are ambiguous in one or both of the manuscripts; here I have chosen the reading most consistent with the other manuscript. The last part of one letter name in Titus is hidden by the present binding, and is shown here as truncated; it might have been identical to the one in Bodleian.

Bodleian			Titus			St. John		Hickes		Stowe
1	a	alar	1	a	alar	1	a	1	a	7	a
2	b	braut	2	b	braut	2	b	2	b	10	b
2	b	braut	2	b	braut	30a	b	30a	b	10	b
3	c	cúsil	3	c	cusil	3	c	3	c
4	d	dexu	4	d	dexu	4	d	4	d
5	e	egui	5	e	egui	5	e	5	e
6	f	fich	6	f	fich	6	f	6	f	1	f
7	g	guichr	7	g	guichr	7	g	7	g
8	h	húil	8	h	huil	8	h	8	h	4	h
9	i	iechuit	9	i	iechu[	9	i	9	i	6	i
10	k	kam	10	k	kam	10	k	10	k
11	l	louber	11	l	louber	11	l	11	l	12	i
12	m	múin	12	m	múin	12	m	12	m	11	m
13	n	nihn	13	n	nihn	13	n	13	n	5	n
14	o	or	14	o	or	14	o	14	o	3	o
15	p	parth	15	p	parth	15	p	15	p
16	q	quith	16	q	quith	16	q	16	q
17	r	rat	17	r	rat	17	r	17	r
18	s	surg	18	s	surg	18a	s	18	s	8	s
18	s	surg	18	s	surg	18b	s	18	s	8	s
19	t	traus	19	t	traus	19	t	19	t	9	t
20	u	úir	20	u	úir	20	u	20	u	2	u
21	x	ieil	21	x	xeil	21	x	21	x
22	y	oyr	22	y	oyr	22	y	22	y	13	y
23	z	zeirc	23	z	zeirc	23	z	23
25	&	estiaul	24		&	24	&	24	&
24	ae	arm	25	ae	arm	25	æ	25	æ
26	eu	egui	26	eu	egui	26	eu	26	eu
27	au	aur	28	au	aur	27	au	27	au
28	ei	einc	29	ei	einc	28	ei	28	ei
29	hinc	hinc	30	hinc	hinc	29	hc̅	29	hunc
30	ego	henc	31	ego	hene	30b	ego	30b	ego
31	ecce	elau	32	ecce	elau	31	ecce	31	ecce
32	uult	utl	33	vult	utl	X	uult	X	uult
33	oe	orn	27	oe	orn

The two N₁ manuscripts both have an oe letter absent from the N₂ group. In Titus, it is in position 27 along with the other diphtongs, but in Bodleian it is in the final position following uult. In the N₂ group, uult has been displaced from the rest of the alphabet, and is located among variant letterforms of an Anglo-Saxon fuþark. Two places before this they both have an œðel-rune which has the epigraphical shape of the rune which the Nemnivan letter seems to be based on, and a value given as oe.

Overall, the four N₁ and N₂ witnesses have rather similar character shapes, but in a few instances there is a clear divergence between the groups, for example in d, f, k, n, p and z. The same is the case for most of the non-alphabetical characters, but note that the au in Hickes has an intermediate form showing that St. John and N₁ are related despite otherwise being dissimilar enough to fall in this cateory. There is also a very probable displacement, where the shape of hinc in N₁ seems to correspond with that of ei in N₂.

Besides containing less than half the number of letters, the Stowe version is more different from each of the others than they are from each other, with several seemingly unrelated letterforms. Still, there are too many clear matches to disregard the relationship. The clearest match is with the a of St. John, followed by its f and, in inverted orientation, its t. The n on the other hand is rather close to Bodleian, where St. John has an entirely different shape. Stowe’s u closely matches the h of all four others. This could very well be due to a reading of the lettername huil as uil at some point in the transmission, leaving the origin of its h a mystery. The shape of its s could easily derive from the q of three of the others, but in this case there is no easily explained source of the confusion.

Section 3, cryptic runes in fuþark order and runic equivalents

This section is introduced by a subheading similar to the previous one, reading alıꞇ̃, normalised aliter meaning otherwise. As this presents rather uncorrupted knowledge of cryptic runes, it would have made much more sense if it immediately followed section 1.

The first part of these entries is a slight variant of twig runes, where large numbers on the right-hand side are split into two smaller groups of twigs, so that the second group branches off from the lowest positioned twig of the first group rather than from the stave. Two runes are erroneously drawn with two twigs where there should only have been one, but otherwise the numbers conform to the established principles of cryptic runes. Two runes, corresponding to m and l breaks with the convention of the majority by starting a second group of twigs after a first group of only two rather than three. It is probably just a coincidence that this splits the third ætt into two along the same lines as the lack of oc before the m-rune divides this ætt in the first section as described above.

To show the value of each cryptic rune, the corresponding normal rune is written after it, except for the first where the letter f is given instead of the f-rune and placed closer to its cryptic rune than the following runic equivalents are to theirs. Just as in section 1, the runes are fairly uncorrupted intermediate type runes. Apart from the missing f-rune, the stave overshoots the bow of the r-rune, making it look like the letter k, and the n-rune has a spurious extra branch on the left-hand side. Unlike in section 1, the o-rune is here the expected intermediate type. The otherwise close match between the two sets of runes strongly suggests that the long-branch form in section 1 very well could have originated as a copying error.

Supplementary runes and other additions in a different ink

Apart from the (at least) two phases of addition already mentioned above, a yet later phase can be identified through stylistic differences and the use of a more brownish ink.

The entry that is placed first in the ordering of the original material is in the o-rune cell in section 1, in the upper right corner like the latter of the two additional runes in the a- and s-rune cells. It provides the intermediate type rune as an alternative, but instead of a ꝉ sign in front of the sign, it repeats the sound value between centered dots after it: ᚮ ·o·. All sound values in this phase are marked this way, while the older material solely uses full stops on the baseline.

The second entry is in the cell of the k-rune (whose value is here given as c), to the right of the sound value like the earlier of the two additional runes in the a- and s-rune cells. It provides not an alternate rune, but an alternate letter equivalent: ꝉ ·q·. The choice of adding this particular value that does not denote any distinction of sound at all comes across as quite ludicrous when the quite frequent g, e and d are not given as alternative values for c, i and t.

This rather odd choice may be due to the author basing the additions on a source where the alternate values were no longer applied to these runes, but supplementary (dotted) runes were used instead. While no such supplemental runes are given for the three mentioned values, ones for p, æ and ø are placed in the first three of the four cells of the table that were not used by the primary material.

Reordered in the same way as the first section of the original content, these three cells read:

	p
	ae	ꝺẏpꞇong̃
	óé	dıpꞇ̃

While the shapes of the three supplementary runes that are included are well preserved, the presentation of the material betrays a complete lack of understanding of it. Including the late and epigraphically rare p while leaving out g and e in particular can only be the result of slavishly copying parts of a list without any familiarity with the usage behind it. The two umlauted vowels are denoted diphtongs. This term does not necessarily have its strict modern meaning, as the Greek digraphs representing long monophtongs are described as diphtongs in the Abecedarium Nord[manniscum], but the accenting of both vowels in the second value suggests that the modern meaning was indeed the intended one. Both instances of the term is abbreviated with slightly ambiguous abbreviation signs simplified to tildes here; the former is presumably intended as the conventional sign for us, while the latter is a general abbreviation sign requiring the meaning to be supplied from context. Following the second instance, there is also a sign with uncertain meaning, but which does not seem to add anything significant.

Both this latest phase and the earlier variant forms look like they have been added by the original author. The letterforms are the same, including the shape of ꝉ, and the first instance of the word diphtong has an instance of non-etymological ẏ for i just like the heading of the alphabet of Aethicus Ister on the following page: Caracteres ęthẏci phẏlos̃, normalised Characteres Aethici philos[ophi], characters of Aethichus [the] philos[opher]. Misunderstanding the umlauted vowel æ as a diphtong should therefore be seen in connection with the lack of the letter Æ in the Anglo-Saxon supplement to the Latin alphabet above.

Tor Gjerde
i@old.no