H and N ablaut

  ILN find much unexpected variation (b/p/p?/m/w; ... 
  s/s./s?/z/d./ etc) in consonants in Semitic roots. Further, he
  finds an impossible number of correspondence sets when trying
  to reconstruct Afro-Asiatic roots (and taking IE roots into
  account).
  He therefore proposes two (pseudo) phonemes, H and N,  which 
  when affixed next to a consonant 'ablauts' (unfortunate choice
  of terminology) that consonant according to the following rules:

  plain      H-ablaut     N-ablaut
  b, p       p? (IE bh)   mb, m
  d, t       t? (IE dh)   nd, n
  g, k       k? (IE gh)   ng, n,
  l          r            n 
  h          £
  s          s?, s.
  ð          d?, d.

  ILN proposes a special relationship between IE and AfroAsiatic 
  within Nostratic (as did Møller). Otherwise there is general
  agreement that AfroAsiatic stands apart from all the other 
  languages within Nostratic. 
  Bomhard notes that IE seems to be related most to Uralic and 
  the other Siberian languages with respect to morphology and 
  inflection, but most to AfroAsiatic (and Kartvelian) with 
  respect to vocabulary. This situation might have come about
  as a result of heavy borrowing by IE from especially Semitic.
  If this is so, then the correspondences of Hodge and the 
  proposed cognates of Møller must be found within those loanwords.