Ortiz-Lira (2000) provides the following example by Halliday  to illustrate the deaccentuation of final relative clauses:
Then he asserts: “We can safely assume that, in English relative clauses, the nuclear accent normally goes on the argument (noun), irrespective of the condition of new or given of the information in the clause, and the length of it”.
The deaccentuation of final relative clauses is no doubt an interesting issue, but it doesn’t appear to be as unproblematic as it sounds in Ortiz-Lira’s statement, which probably needs some qualification. There are two elements missing in it which should be taken into account.
Firstly, this tendency only applies to defining relative clauses. Non-defining relative clauses are parenthetical and normally have an IP of their own. As in the example provided by Carr (2013):
Secondly, relative clauses tend to be deaccented when the antecedent is the object of the verb. If it functions as subject, the deaccentuation seems really dubious and by no means can be considered the unmarked option.
Even if the verb in the embedded clause is intransitive and consequently there is no final object to take the tonic, there would probably be no deaccentuation, as in:
(190) *Sewing is a job that barely pays.
When the antecedent is the object, instead, the relative clause is left out of focus.
It is also interesting to note that the length of Hallidays’ example has nothing to do with the relative clause per se. It is mainly due to the addition of a long time adverbial at the end of the utterance, which are usually deaccented , an also to the fact that the noun phrase tape recorded is an open compound .
So, it looks like the accentual pattern dealt with here has close connections with the cases studied in 3.6-3.9, where always a syntactic movement is involved. So to avoid confusion when this accentual pattern is taught, it would be important to make these two conditions for the deaccentuation of relative clauses very clear.
This notion seems to be borne out by the fact that in all the examples provided by Ortiz-Lira and Wells about this point the antecedent is the object and never the subject.
This warning expressed, it must be said that deaccented final non-defining relative clauses (in which the antecedent is the object) are absolutely commonplace. In fact, it is highly advisable to encourage the use of this intonational pattern in EFL, since it will certainly make the English of Spanish learners sound more natural.
The following examples, taken from different sources, support the view expressed above.
(194) He was just trying to protect you from all the naughty things we were doing (unknown child actor).
(195) You might be interested in this tie that I’m wearing (Hugh Laurie; Oxford, UK).
(196) And, if you’re unhappy at the end of the year, I’ll buy out and you can go back to that cozy two percent your bank is giving you (David Hyde Pierce; New York, US).
 Halliday, M.A.K. (1970) A course in spoken English: intonation. London: Oxford University Press.