Greetings.

> There are some misprints or mistakes in the book, and one whole class
> of them involves Old Icelandic. Can anybody tell me what these forms
> represent? Oldest attestations, or maybe a mnemonic scheme for learning
> the paradigms?

These are merely alternative forms. In the mss we have both /e/ and /i/
for endings such as masculine dative and both /o/ and /u/ for endings
such as u-stem masculine accusative.

> I have Gordon's book, which I take as my authority.
>
> 1. Lass has /e/ for /i/ wherever this appears in the dative singular
> masc and neut. Even /deg-e/, where he explains the stem change is due
> to i-umlaut. For i-stem gestr Lass gives nom. pl. gest-er and acc. pl.
> gest-e.

Perfectly normal. The i-mutation in /degi/ is actually somewhat more
involved - but it's certainly i-mutation.

> 2. The u-stem example is even stranger:
> sg.
> N sun-r for son-r
> G son-ar
> D syn-e for syn-i
> A sun for son
> pl.
> N syn-er for syn-ir
> G son-a
> D sun-um
> A sun-o for sun-u

He seems to be going for the most archaic forms. You notice he only
uses the a-mutated forms where there is an /a/ in the ending. This is
rather theoretical and hardly well attested in the mss.

> BUT a dative -i for foet-i (with oe=ligature)

That is rather peculiar, I admit. Usually people are consistent in
their choices of /e/ vs. /i/ and /o/ vs. /u/.

> 3. Here is the present conjugation of bera
> sg
> 1 bær-a
> 2 bær-er
> 3 bær-e
> pl
> 1 bær-em
> 2 bær-eÞ
> 3 bær-e

This is subjunctive past tense and perfectly correct.
Another choice in this author's spelling seems to be
to use thorn exclusively rather than alternating eth
and thorn. This, too, is perfectly permissible.
(Note that you seem to be mistaking 'Þ' for small thorn.)

> I understand from Gordon that ö (hook o) to á and then i-ulaut to æ.
> So why /e/ for /i/ in all the endings?
>
> 4. Last thing: for '9' and '10' Lass gives OIc nió and tió, instead of
> the expected níu and tíu; '11' is ellefo for ellifu.

If he really has the accent on the o's I don't know why. I could understand
using no accents on those words, though. That would be consistent with what
we know of Old Norse phonology.

In summary this author seems to be going for a spelling closer to the
oldest extant Icelandic manuscript than most of us here are using.

Kveðja,
Haukur