This was a difficult one to work out, but I think I've got it now. What I was unsure of was what "marcas" referred to, since "marker" is a morphological term (as in "the plural marker -s" in English), and so "active marker" would mean a marker which signals the active voice - but this is about phonology and nativisation (plus "marker" was "marcador" in one of your other questions, not "marca").

After some hunting, I came across this:

"Subjacentemente, alguns segmentos não têm toda a informação fonológica (tds os traços). Estes segmentos são, geralmente, ‘menos marcados’
***Noção de ‘marca’*** relaciona-se com a frequência de ocorrência de determinados segmentos nas línguas ou numa língua e com os processos fonológicos a que estão sujeitos."
http://209.85.135.104/search?q=cache:TS8EEEcstgkJ:w3.ualg.pt...

This is called "markedness" in phonology:

"The **notion of markedness** was first developed in Prague school phonology... When two phonemes are distinguished by the presence or absence of a single distinctive feature, one of them is said to be marked and the other unmarked for the feature in question. For example, /b/ is marked and /p/ unmarked with respect to voicing."
http://www.britannica.com/eb/article-35106/linguistics

"On the basis of typological facts, onset clusters with a sonority rise (such as /bn/, /b/ being less sonorous than /n/) are said to be universally less marked than those with a sonority plateau (such as /bd/), which are in turn less marked than those with a sonority fall (such as /lb/). Using three different tasks – syllable judgment, discrimination, and lexical decision – the authors show that the more an onset cluster is marked, the more likely it is perceived with an illusory epenthetic vowel by English listeners."
http://www.ehess.fr/centres/lscp/persons/peperkamp/Peperkamp...

Certain distinctive features are said to be "marked" or "unmarked" in a given language. In the example above, "marked" features or segments are those which a language regards as less usual or desirable, whereas "unmarked" ones are "the norm", as it were. For each language we can postulate a specification for each feature as to whether it is marked or unmarked. These specifications collectively are the information relied on by a listener in the process of deciding whether or not a given input (foreign sound) conforms or not to the phonology of his/her native language, and so marked segments/features will tend to be blocked.

"Notice also that the SBG account produces representations which preserve **markedness information**. The right metrical bracket is preserved in the output, indicating the lexically marked positions."
http://216.239.59.104/search?q=cache:V0C7XC_X1pEJ:society.ki...

"Candidate chains and opacity (somewhat simplified)
•Interaction of raising and epenthesis in Bedouin Arabic
-Visibly **active markedness*** constraints:
*COMPLEX-CODA — violated by final cluster in *[gabr].
RAISE — violated by any [a] in a nonfinal open syllable, such as [ga,bur]."
http://socrates.berkeley.edu/~ling298/roa_823.pdf

--------------------------------------------------
Note added at 5 days (2007-11-19 19:36:41 GMT)
--------------------------------------------------

Yes, you're right - in that last link I really ought to have put the stars around the whole phrase, not just the two words. The structure of the phrase is thus [active [markedness constraints]] rather than [[active markedness] constraints].