Staffan Fridell, from Uppsala's Nordic department, gave a talk today about the different ways of sorting and organizing dental fricatives in Old Swedish. Traditionally, the voiced dental fricative [ð] and the voiced alveolar plosive [d] have been thought of as allophones (that is to say, variants within a type) of the /ð/ phoneme. This viewpoint is put forth by Bengt Pamp in his 1971 book Svensk språk- och stilhistoria, one of the first attempts to comprehensively describe Swedish phonetics between 1200 and 1500. Tomas Riad's 2002 volume The Nordic Languages follows in this vein.

Fridell is interested in exploring the possibility that the voiced dental fricative [ð] and the voiceless dental fricative /θ/ (often written /þ/ in Scandinavian research) are actually allophones of the voiceless /θ/ phoneme. Some compelling evidence for this alignment is the way that it fits in with the other voice and unvoiced allophones which various runes stood for:

Rune	Phoneme	voiced	unvoiced
	þ	ð	þ
	t	d	t
	k	g	k
	b	b	p
	f	v	f
	h	ɣ	x

Yet the above complimentary distribution is not the main thrust of Fridell's argument; rather, he presents the relationship between phoneme (sound) and grapheme (written character) in both Runic and Latin manuscripts as his main justification for seeing [ð] and [þ] as related. He presented some interesting preliminary results, drawn from un-normalized primary texts, that show which graphemes people used to record [ð] and [þ].

Fridell sees "a change in the phonetic belonging of [ð] in that [þ] becomes voiced in medial and final position during the late proto-nordic period." Later a different change in the status of [ð] happens when initial [þ] changes to the stop [t] around 1400 AD. At that point [ð] returns to its previous status of allophone to [d].