Fig. 1

Experiment 1 design. The participants heard sequences of number words in Hebrew, and repeated each sequence. Each of the 4 conditions included 20 stimuli (sequences), derived from the same 20 numbers (in this example: 65,432). The stimulus words were shuffled in a different manner in each condition, according to that condition’s fixed word order (shown here on the right). Thus, the number and length of grammatical segments was fixed for each condition, and different between conditions. A stimulus included either 6 words (as in this example) or 7 words, depending on the participant’s short-term memory capacity. For the 7-word stimuli, the word order in each condition was A [hundreds, tens, ones, thousand, hundreds, tens, ones]; B [thousand, hundreds, tens, ones], [hundreds, tens, ones]; C [hundreds, tens], [hundreds, ones], [tens, ones, thousand]; D [thousand, ones], [ones], [tens], [tens], [hundreds], [hundreds]