Towards a distributed connectionist account of cognates and interlingual homographs: Evidence from semantic relatedness tasks [Experiment 2]

These materials were used for Experiment 2 of Poort and Rodd's (2019) paper "Towards a distributed connectionist account of cognates and interlingual homographs: Evidence from semantic relatedness tasks". The main aim was to investigate whether a single encounter with a cognate or interlingual homograph in one's native language influences subsequent processing of these words in one's second language.

From Poort and Rodd (2019): "Experiment 2 had two aims. First, given the surprising lack of a cognate facilitation effect in Experiment 1, to conduct a direct replication of the pattern of results observed in the bilingual group of Experiment 1. And second, given the successful use of a semantic relatedness task in the monolingual domain to replicate effects of long-term priming with homonyms (Gilbert et al., 2018), to attempt once again to replicate the cross-lingual long-term priming effect demonstrated by Poort et al. (2016).

...

[The] participants completed a self-report language background survey in Dutch at the start of the experiment to verify eligibility (see the Participants section for the eligibility criteria). The experiment comprised five separate tasks: (1) the Dutch version of the LexTALE (Lemhöfer & Broersma, 2012), (2) the Dutch semantic relatedness task (mean duration in mm:ss: 08:38), (3) the Towers of Hanoi task (with instructions presented in English; maximum duration set to 4 minutes), (4) the English semantic relatedness task (mean duration: 10:53) and (5) the English version of the LexTALE (Lemhöfer & Broersma, 2012). Across participants, English semantic relatedness judgements to primed items were made on average 14 minutes and 27 seconds after they were primed in the Dutch semantic relatedness task (SD = 01:51; range = 11:37-25:36). The five tasks were presented separately with no indication that they were linked. "

The stimuli and processing scripts are also available on the Open Science Framework: https://osf.io/2at49/.

Back to Open Materials


Complete experiment

This is the complete experiment.

Note: parts of this experiment (the consent form, demographics questionnaire, Dutch LexTALE, Dutch semantic relatedness task and debrief) are in Dutch. Branching prevented people who were not eligible to participate from continuing past the demographics questionnaire. To get past this while previewing, please select the following answers for the following questions:

  • "Wat is je leeftijd?": Any value between 18 and 50 (inclusive)
  • "Heb je dyslexie en/of een andere taalstoornis?": "Nee."
  • "Wat is je nationaliteit?": "Nederlands." or "Belgisch."
  • "Waar woon je nu?": "Nederland." or "België."
  • "Spreek je Nederlands/Vlaams vloeiend?": "Ja."
  • "Spreek je Engels vloeiend?": "Ja."
  • "Is Nederlands/Vlaams (één van) je moedertaal/-talen?": "Ja."

Creative Commons Attribution (CC BY)


Dutch semantic relatedness task

The Dutch semantic relatedness task was the priming task. Participants read sentences in Dutch which contained either a cognate, interlingual homograph or the Dutch translation of an English control word. Each sentence was followed by a probe and participants were asked to indicated whether the probe was semantically related to the sentence or not.

From Poort and Rodd (2019): "This task served to prime the cognates, interlingual homographs and translation equivalents. To ensure the participants processed the prime sentences, participants were asked to indicate via button presses whether a subsequent probe was semantically related to the preceding sentence. The 50 target sentences for each of the three word types were pseudorandomly divided into two sets of 75, matched for all key variables and prime sentence length, for use in the two versions of the experiment. Including the 24 primed filler items, participants read a total of 101 sentences. Half of the sentences in each version were paired with related probes and half with unrelated probes. A practice block of six sentences was followed by four blocks of 24 or 25 sentences (mixed targets and fillers). The order of the items within a block was randomised for each participant, as was the order of the blocks. Of all the items of each word type that were assigned to a block (6 or 7 items), 1 (or 2) item(s) was/were assigned to each of the six blocks of the English semantic relatedness task. This was to ensure that the (variation in) duration of the delay would be similar for each of the three word types. A 15-second break was enforced after each block and three fillers (additional to the 24 cognate, interlingual homograph and translation equivalent fillers) were presented at the start of each block. Participants were allowed to read the sentences at their own pace, with a minimum presentation time of 1000 ms and a maximum of 4000 ms. Participants pressed the spacebar after they had read the sentence, after which the probe appeared on the screen. Each probe remained on the screen until the participant responded or until 2000ms passed. The inter-trial interval was 500ms."

Explanation of the different trial set-ups:

  • instructions: presents instructions for the task.
  • practice: trial set-up for a practice trial (with feedback on speed and accuracy performance).
  • continue: briefly repeats the instructions again after the practice block.
  • filler: trial set-up for the first three trials of each block. These items were lead-in fillers which we did not include in the analyses. This is the same trial set-up as "main", but giving it a different name in Gorilla allowed us to filter these trials out of the data more easily.
  • main: trial set-up for an experimental item.
  • break: 15-second break.

There are three spreadsheets. The spreadsheet "debug" contains only a few items and was used for testing the task. The spreadsheets called "v1" and "v2" contain the items for version 1 and version 2, respectively. From Poort and Rodd (2019): "Priming was a within-participants/within-items factor: for each participant, half the targets and eight of the fillers of each word type were primed (i.e. appeared during the priming phase) while half were unprimed (i.e. only occurred in the test phase). Two versions of the experiment were created such that participants saw each experimental item only once but across participants items occurred in both the primed and unprimed conditions. The same fillers were primed for all participants."

Creative Commons Attribution (CC BY)


The Towers of Hanoi

This implementaiton of the Towers of Hanoi task was created by Gorilla and is available in the Samples Library.

From Poort and Rodd (2019):

"This task served to introduce a delay between priming and testing, while minimising exposure to additional linguistic material. The Towers of Hanoi is a puzzle in which disks of progressively smaller sizes must be moved from one peg to another. There are two simple rules: (1) only one disk may be moved at a time and (2) a larger disk may not be placed on top of a smaller disk. The goal is to move the disks from the starting peg to the finish peg in as few moves as possible. Participants were given 4 minutes to complete as many puzzles as they could, starting with a puzzle with three disks and three pegs. Each subsequent puzzle had the same number of pegs but one disk more than the previous puzzle. The instructions were presented in English to minimise any general language switching cost on the English semantic relatedness task."

Creative Commons Attribution (CC BY)


Gorilla Experiment Builder
http://gorilla.sc/


English semantic relatedness task

The English semantic relatedness task was the main task of the experiment. Participants were presented with two words, one after the other, and asked to indicate whether the two words were related to each other not. Half of the target items (the first word of the "yes"-trials) that participants saw in this task were words (whose translations) they had encountered during the Dutch semantic relatedness task.

This task is the exact same task as the English semantic relatedness task for Experiment 1, except that participants were given only 2000 ms to respond, to reduce the overall duration of this task and of the whole experiment. A warning was presented to the participant that they were responding too slowly if they had not responded 1500 ms after the probe first appeared. The warning remained on screen for 500 ms, during which time the participant could still respond.

Description of the task from Poort and Rodd (2019) for Experiment 1: "... During the semantic relatedness task, the participants saw all 150 related target–probe pairs (“yes”-responses) and all 150 unrelated filler–probe pairs (“no”-responses) and were asked to indicate, by means of a button press, as quickly and accurately as possible, whether the word they saw first was related in meaning to the word they saw second. A practice block of 6 pairs was followed by 6 blocks of 50 experimental pairs. The order of the pairs within blocks was randomised for each participant, as was the order of the blocks. As a lead in, three filler pairs were presented at the beginning of each block (these were additional to the 150 filler pairs described above), with a 15-second break after each block. The target or filler item would appear first on screen and remain for 200 ms. After a blank screen lasting 50 ms, the probe appeared. ..."

Explanation of the different trial set-ups:

  • instructions: presents instructions for the task.
  • practice: trial set-up for a practice trial (with feedback on speed and accuracy performance).
  • continue: briefly repeats the instructions again after the practice block.
  • filler: trial set-up for the first three trials of each block. These items were lead-in fillers which we did not include in the analyses. This is the same trial set-up as "main", but giving it a different name in Gorilla allowed us to filter these trials out of the data more easily.
  • main: trial set-up for an experimental item.
  • break: 15-second break.

Creative Commons Attribution (CC BY)


English LexTALE

This is a speeded version of Lemhöfer and Broersma's (2012) English LexTALE task. The stimuli and instructions were taken from Lemhöfer and Broersma's (2012), but in keeping with the other task in the experiment, which was also speeded, a 2000 ms time limit for responding was added.

From the LexTALE website: "The LexTALE is a quick and practically feasible test of vocabulary knowledge for medium to highly proficient speakers of English as a second language. It consists of a simple un-speeded[1] visual lexical decision task. In contrast to other vocabulary or proficiency tests, it has been designed to meet the needs of cognitive researchers. It is quick, easy to administer, and free, and yet it is a valid and standardized test of vocabulary knowledge. It has also been shown to give a fair indication of general English proficiency."

[1] But note that we added a 2000 ms time limit.

Gorilla Open Materials Attribution-NonCommerical Research-Only


Lemhöfer, K. & Broersma, M. Behavior Research Methods (2012) 44(2), 325-343. DOI: 10.3758/s13428-011-0146-0.
www.lextale.com


Dutch LexTALE

This is a speeded version of Lemhöfer and Broersma's (2012) Dutch LexTALE task. The stimuli and instructions were taken from Lemhöfer and Broersma's (2012), but in keeping with the other task in the experiment, which was also speeded, a 2000 ms time limit for responding was added.

From the LexTALE website: "The LexTALE is a quick and practically feasible test of vocabulary knowledge for medium to highly proficient speakers of English as a second language. It consists of a simple un-speeded[1] visual lexical decision task. In contrast to other vocabulary or proficiency tests, it has been designed to meet the needs of cognitive researchers. It is quick, easy to administer, and free, and yet it is a valid and standardized test of vocabulary knowledge. It has also been shown to give a fair indication of general English proficiency."

[1] But note that we added a 2000 ms time limit.

Gorilla Open Materials Attribution-NonCommerical Research-Only


Lemhöfer, K. & Broersma, M. Behavior Research Methods (2012) 44(2), 325-343. DOI: 10.3758/s13428-011-0146-0.
www.lextale.com

Preferred Citation Poort ED, Rodd JM. 2019. Towards a distributed connectionist account of cognates and interlingual homographs: Evidence from semantic relatedness tasks. PeerJ 7:e6725.
http://doi.org/10.7717/peerj.6725
Poort, E. D., & Rodd, J. M. (2019, February 26). Towards a distributed connectionist account of cognates and interlingual homographs: Evidence from semantic relatedness tasks.
http://psyarxiv.com/fe2k8
Conducted at University College London
Published on 15 April 2019
Corresponding author Dr Eva Poort Experimental Psychology
University College London

eva.poort.12@ucl.ac.uk