Available at

5 Conclusion 
This thesis first aimed at finding out what the most challenging linguistic item was 
for interpreter trainees. It also focused on investigating students’ interpreting 
quality to see if there was any marked improvement in the rendition of chosen 
linguistic items across time. Complex noun phrases turn out to be the most 
challenging linguistic item for interpreter trainees as only over half of all NPs were 
correctly translated. The “N + subclause” sub-pattern is the most difficult among 
the three investigated sub-patterns as it has a very low percentage of close 
renditions (45.28%) and a high percentage of divergent renditions (24.42%).
In order to answer the second research question (improvement in quality across 
time), I focused on five students as they carried out at least three interpreting tasks 
(STU01, STU02, STU04, ST06 and STU07) and on their proportion of close and 
divergent renditions across the interpreting tasks. Unfortunately, there is no clear 
trend and no marked improvement in the interpreting quality is to be observed 
across time. Global features can be linked to students’ performances, but these 
observations would need more in-depth investigation. First, it seems that when the 
student’s delivery rate is much lower than the corresponding IN delivery rate, the 
student correctly translates less than half of the occurrences. There also seems to 
have a direct link between a faster delivery rate and a higher proportion of 
divergent renditions. As for lexical density, STU lexical density scores and STU 
divergent renditions could be related. Indeed, there are four students who support 
this idea, as the IN for which their LD score is the lowest is also the IN for which 
their percentage of divergent renditions is the highest.
This study has of course its own limitations. It is first really important to keep in 
mind that this is only one way to assess the quality of an interpretation and does 
not mean that they did not improve across time at all. Results could also have 
been different if numbers were more complex for example. This study only focuses 
on different linguistic items, but my database also comprises the students’ 

---

Conclusion 
 
page 94 
disfluencies (repetitions, hesitations, self-corrections, silent pauses). These 
disfluency markers have not been investigated in the present study due to a lack 
of time, but it would be very interesting to link these to students’ performances in a 
further research. 
Speaking of future research, we have seen that complex noun phrases are the 
most challenging items for interpreter trainees and we investigated three types of 
NPs as they were the most frequent in the database: “A + N”, “N + prep + N”, and 
“N + subclause”’. However, it would be of a great interest to carry out a research 
on other types of NPs, such as “N + N” noun phrases, as this pattern is much more 
frequent in English than in French (Paillard, 2000: 49-51). It seems that this type 
of NPs can be very difficult for interpreter trainees as Lefer & De Clerck (2019) 
state that more than half of their “N + N” noun phrases had been simplified by 
professional interpreters. In addition, investigating other language pairs can also 
be very meaningful. It will then be possible to state if the source language, here 
English, has the same influence on interpreter trainees having different mother 
tongues (Italian or Spanish for example). 

---

Appendices 
 
page 95