Propose two analyses for given data: an abstract analysis, with abstract underlying segments and a concrete analysis with exception features.

[sen't-amos] we sit / ['sjent-o] I sit

[kon't-amos] we count / ['kwent-o] I count

[ben'd-amos] we sell / ['bend-o] I sell

[mon'tamos] we mount / ['mont-o] I mount

[alje'n-amos] we alienate / [a'ljen-o]I alienate

[frekwen't-amos] we frequent / [fre'kwent-o] I frequent

So far I have these things:

surface alteration / underlying form

non-alternation [e] or [o] / [e] or [o]

non-alternation [je] or [we] / [je] or [we]

[e] or [o] alternating with [je] or [we]/ ?

we have these alternations under the stress situation.

Now I have a question are these analyses correct? what are abstract underlying segments? how about concrete analysis with exception features?

  • What problem would that be? Are you supposed to write rules to delete word-final mos?
    – user6726
    Apr 16, 2016 at 19:38
  • actually I mentioned these final words just because of the stress issue. The problem is I can not write concrete analysis with exception features. and also what are abstract underlying segments?
    – liza
    Apr 16, 2016 at 19:48
  • 1
    An abstract segment is a segment that does not appear phonetically in any instantiation of the morpheme. A so-called "concrete" analysis assumes underlying forms that are either composed of segments that appear in some surface allomorph of the morpheme, or some actual surface allomorph. A completely concrete analysis does not distinguish underlying and surface representations.
    – user6726
    Apr 16, 2016 at 19:54
  • would you please give me one examle of "abstract underlying segment" of above date?
    – liza
    Apr 16, 2016 at 20:20
  • For example frekwento ← /freqento/. Or, ← /frekænto/. There are very many abstract analyses.
    – user6726
    Apr 16, 2016 at 21:15

1 Answer 1


I hope I am answering the question.

When Spanish was evolving very early (600AD plus or minus 400 years), a process called "vowel breaking" occured, which basically turned some [o] into [we], and some [e] into [je], when they are stressed. The evolution is something like as follows:

  • Latin: computáre, cómputo (accent for stress)
  • [kombu'tare], ['kombuto]
  • [komb'tar], ['kombto]
  • [kom'tar], ['kwemto]
  • [kon'tar], ['kwento]
  • Modern Spanish: contar, cuento

Since it happened too early, modern speakers would not have ['konto] as the underlying form of ['kwento].

However, if you had to impose an underlying form, it would be ['konto].

In general, some [e] and [o] in stressed positions change to [je] and [we] respectively.

abstract analysis (my attempt)

Some [e] and [o] change to [je] and [we] when stressed. Some don't change when stressed. Some are [je] and [we] overall.

The [e] and [o] that change must have a different underlying representation than those that do not change. Perhaps those that change are lower than others, i.e. [ɛ] and [ɔ].

Moreover, those [je] and [we] can keep those as their underlying representations.

Underlying representation | surface form
[sɛn't-amos]              | [sen't-amos]
['sɛnt-o]                 | ['sjent-o]
[kɔn't-amos]              | [kon't-amos]
['kɔnt-o]                 | ['kwent-o]
[ben'd-amos]              | [ben'd-amos]
['bend-o]                 | ['bend-o]
[mon'tamos]               | [mon'tamos]
['mont-o]                 | ['mont-o]
[alje'n-amos]             | [alje'n-amos]
[a'ljen-o]                | [a'ljen-o]
[frekwen't-amos]          | [frekwen't-amos]
[fre'kwent-o]             | [fre'kwent-o]

concrete analysis (my attempt)

[e] and [o] when stressed becomes [je] and [we] except monto and vendo.

  • You might want to consider a phenomenon known as rule-inversion where earlier a → b / __ c becomes b → a / __ ~c.
    – user6726
    Apr 17, 2016 at 13:55
  • Please do refine my answer by editing it directly.
    – Kenny Lau
    Apr 17, 2016 at 13:58

Not the answer you're looking for? Browse other questions tagged or ask your own question.