Table 1.

Parameters across linguistic levels. Number of elements considered (N), mean, standard deviation (s.d.), mode, median and percentiles 10 (p10) and 90 (p90) of a physical magnitude (time duration distribution) versus symbolic ones (number of characters, number of phonemes and number of words) for the three linguistic levels (phoneme, words and BGs). Since speakers sometimes omit or add phonemes to the same word, the number of characters per phoneme is obtained indirectly averaging number of phonemes and number of characters in the word. The p10 and p90 percentiles give us an account of the range of durations, without considering outliers.

	time duration t (seconds)
	N	mean 〈t〉	s.d.	mode	median	p10	p90
phoneme	8 × 105	0.08	0.06	0.05	0.07	0.03	0.14
words	3 × 105	0.24	0.17	0.12	0.2	0.08	0.45
BG	5 × 104	1.4	1.2	0.4	1.1	0.3	3.1
		mean	s.d.	mode	median	p10	p90
number of characters
phoneme		1.4	0.5	1	1.3	2	2
words		4	2	4	4	2	7
BG		24	23	2	17	3	54
number of phonemes
words		3	1.6	2	3	1	5
BG		18	17	2	13	2	40
number of words
BG		6	6	1	4	1	13