Investigating the Effects of Valence, Arousal, Concreteness, and Humor on Words Unique to Singapore English

HIGHEST			LOWEST
WORD	M	SD	WORD	M	SD
Valence
ang bao/hong bao	8.29	1.07	chee ko pek/ti ko pek	2.33	1.60
huat	8.00	1.31	Ah Long	2.61	1.52
shiok	7.88	1.43	saman	2.80	1.54
makan	7.87	1.42	O$P$	3.11	2.16
pasar malam	7.54	1.26	hao lian	3.14	1.55
Arousal
bo ta bo lan pa	7.12	1.58	lepak	3.51	2.18
chee bai	7.01	1.78	meh	3.59	1.92
huat	6.99	1.84	abit	3.69	1.82
si mi lan jiao	6.85	1.91	helication	3.92	1.73
hong gan	6.83	1.83	handphone	4.01	1.97
Concreteness
kopi	4.68	0.78	mah	1.41	0.74
handphone	4.65	0.83	ah	1.52	0.97
ang bao/hong bao	4.56	0.88	lah	1.52	0.94
kopitiam	4.52	1.00	ar	1.53	0.88
makan	4.52	0.90	nia	1.56	0.86
Humor
your head	3.95	0.97	handphone	1.56	0.85
pak chiu cheng	3.89	1.18	kopitiam	1.77	0.88
sod	3.86	0.88	kampung	1.93	0.91
bo geh	3.85	1.10	da bao	2.00	0.98
piak piak	3.84	1.27	kopi	2.05	1.03

Table 2

Correlation of lexical-semantic and affective measures for 282 Singapore English items.

	VALENCE	AROUSAL	CONCRETENESS	HUMOR
Valence	1	–0.34***	0.12*	–0.41***
Arousal	–0.34***	1	–0.01	0.55***
Concreteness	0.12*	–0.01	1	–0.26***
Humor	–0.41***	0.55***	–0.26***	1

[i] * p < .05, *** p < .001.

Density plots for valence, arousal, concreteness, and humor ratings provided by human raters.

Table 3

Descriptive information for raw human ratings.

RATING	MEAN	SD	MEDIAN	MIN	MAX	RANGE	SKEW	KURTOSIS
Valence	4.94	1.04	4.75	2.33	8.29	5.96	0.62	0.26
Arousal	5.42	0.68	5.45	3.51	7.12	3.61	–0.19	–0.18
Concreteness	2.83	0.73	2.72	1.41	4.68	3.27	0.47	–0.38
Humor	3.00	0.42	3.03	1.56	3.95	2.39	–0.35	0.13

Table 4

Correlations of human ratings with ChatGPT ratings.

MEASURE	GENERAL (RAW)	GENERAL (WEIGHTED)	SPECIFIC (RAW)	SPECIFIC (WEIGHTED)
Valence	0.40	0.42	0.76	0.78
Arousal	0.26	0.27	0.57	0.59
Concreteness	0.29	0.31	0.66	0.69
Humor	0.18	0.19	0.33	0.39

[i] All correlations were statistically significant, all ps < .01.

Density plots for valence, arousal, concreteness, and humor ratings generated by GPT4 (specific, weighted condition) overlaid on the corresponding human ratings.

Table 5

Descriptive information for LLM ratings (specific, weighted).

RATING	MEAN	SD	MEDIAN	MIN	MAX	RANGE	SKEW	KURTOSIS
Valence	5.30	1.55	5.08	1.00	9.00	8.00	–0.19	–0.01
Arousal	4.36	1.45	4.33	1.03	8.07	7.04	0.51	–0.50
Concreteness	2.40	1.09	2.06	1.00	5.00	4.00	1.12	0.70
Humor	2.93	0.40	3.00	1.00	4.17	3.17	–0.78	5.06

Table 6

Descriptive statistics for 135 word items.

STATISTIC	N	MEAN	ST. DEV.	MIN	MAX
no. of letters	135	4.87	1.67	2	12
orthographic neighborhood size	135	4.01	5.74	0	26
mean bigram frequency	135	2,907.95	1,485.98	226.00	6,993.67
log frequency	135	2.80	1.60	0.00	6.76
valence	135	0.18	0.61	–1.26	1.90
arousal	135	–0.13	0.36	–1.00	1.00
concreteness	135	–0.02	0.74	–1.51	1.92
humor	135	–0.13	0.43	–1.64	0.80

Table 7

Correlation of lexical-semantic and affective measures for 135 word items.

	NO. OF LETTERS	ORTHOGRAPHIC NEIGHBORHOOD SIZE	MEAN BIGRAM FREQUENCY	LOG FREQUENCY	VALENCE	AROUSAL	CONCRETENESS	HUMOR
no. of letters	1	–0.62***	0.21*	–0.44***	–0.11	–0.03	0.30***	0.06
orthographic neighborhood size	–0.62***	1	0.07	0.39***	0.10	–0.12	–0.29***	–0.04
mean bigram frequency	0.21*	0.07	1	–0.03	–0.02	–0.08	0.11	–0.01
log frequency	–0.44***	0.39***	–0.03	1	0.35***	–0.18*	–0.10	–0.29***
valence	–0.11	0.10	–0.02	0.35***	1	–0.28***	0.08	–0.34***
arousal	–0.03	–0.12	–0.08	–0.18	–0.28***	1	0.001	0.52***
concreteness	0.30***	–0.29***	0.11	–0.10	0.08	0.001	1	–0.25**
humor	0.06	–0.04	–0.01	–0.29	–0.34***	0.52***	–0.25**	1

[i] * p < .05, ** p < .01, *** p < .001.

Table 8

Visual Lexical Decision Models.

	DEPENDENT VARIABLE
	RT		ACC
	LINEAR MIXED-EFFECTS		GENERALIZED LINEAR MIXED-EFFECTS
	BASE	BASE+NORMS	BASE	BASE+NORMS
	(1)	(2)	(3)	(4)
no. of letters	54.121*** (11.434)	52.229*** (11.246)	0.006 (0.170)	0.041 (0.156)
orthographic neighborhood size	64.943*** (12.387)	56.943*** (12.024)	–0.473** (0.167)	–0.397** (0.152)
mean bigram frequency	2.057 (9.596)	3.701 (9.181)	–0.265* (0.131)	–0.249* (0.118)
log frequency	–88.212*** (9.943)	–93.193*** (10.627)	1.664*** (0.147)	1.668*** (0.146)
valence		–20.299* (10.013)		0.456*** (0.132)
arousal		–27.718* (11.355)		0.487*** (0.144)
concreteness		–24.448* (10.405)		0.211 (0.128)
humor		–13.877 (11.662)		0.192 (0.146)
Constant	779.558*** (17.712)	776.354*** (17.488)	2.028*** (0.185)	2.023*** (0.177)
Observations	5,554	5,554	7,510	7,510
Log Likelihood	–37,982.720	–37,960.160	–2,684.140	–2,670.626
Akaike Inf. Crit.	75,981.440	75,944.310	5,382.280	5,363.251
Bayesian Inf. Crit.	76,034.420	76,023.780	5,430.748	5,439.415

[i] Note: **p < 0.05; **p < 0.01; ***p < 0.001.

Table 9

Comparing human-generated and LLM-generated norms.

	DEPENDENT VARIABLE
	RT		ACC
	LINEAR MIXED-EFFECTS		GENERALIZED LINEAR MIXED-EFFECTS
	HUMAN	CHATGPT	HUMAN	CHATGPT
	(1)	(2)	(3)	(4)
no. of letters	52.229*** (11.246)	56.792*** (11.256)	0.041 (0.156)	0.009 (0.164)
orthographic neighborhood size	56.943*** (12.024)	58.536*** (12.161)	–0.397** (0.152)	–0.366* (0.162)
mean bigram frequency	3.701 (9.181)	–2.214 (9.470)	–0.249* (0.118)	–0.236 (0.128)
log frequency	–93.193*** (10.627)	–90.895*** (9.935)	1.668*** (0.146)	1.685*** (0.146)
valence	–20.299* (10.013)		0.456*** (0.132)
arousal	–27.718* (11.355)		0.487*** (0.144)
concreteness	–24.448* (10.405)		0.211 (0.128)
humor	–13.877 (11.662)		0.192 (0.146)
valence (gpt)		–4.719 (9.280)		0.137 (0.126)
arousal (gpt)		–10.646 (9.312)		0.245* (0.123)
concreteness (gpt)		–22.781* (9.747)		0.165 (0.129)
humor (gpt)		–32.490*** (9.469)		0.358** (0.129)
Constant	776.354*** (17.488)	777.769*** (17.509)	2.023*** (0.177)	2.031*** (0.181)
Observations	5,554	5,554	7,510	7,510
Log Likelihood	–37,960.160	–37,961.750	–2,670.626	–2,677.244
Akaike Inf. Crit.	75,944.310	75,947.500	5,363.251	5,376.488
Bayesian Inf. Crit.	76,023.780	76,026.970	5,439.415	5,452.652

[i] Note: **p < 0.05; **p < 0.01; ***p < 0.001.

References

Authors

Metrics

Articles in this issue

DOI: https://doi.org/10.5334/joc.470 | Journal eISSN: 2514-4820

Journal RSS Feed

Language: English

Submitted on: Apr 6, 2025

Accepted on: Oct 19, 2025

Published on: Nov 7, 2025

Published by: Ubiquity Press

In partnership with: Paradigm Publishing Services

Publication frequency: 1 issue per year

Keywords:

lexical-semantic norms,

affective norms,

valence,

arousal,

concreteness,

humor,

Singapore English,

visual lexical decision

© 2025 Cynthia S. Q. Siew, Feria Chang, Jin Jye Wong, published by Ubiquity Press
This work is licensed under the Creative Commons Attribution 4.0 License.

Volume 8 (2025): Issue 1

Investigating the Effects of Valence, Arousal, Concreteness, and Humor on Words Unique to Singapore English

Figures & Tables

Table 1

Table 2

Figure 1