Journal of Child Language Acquisition and Development – JCLAD
Vol: 10 Issue: 3 572-600, 2022
ISSN: 2148-1997
572
Children’s humour development: A linguistic perspective
Giulia Baker1
Freelance researcher & Teacher
Michelle Aldridge2
Cardiff University
Abstract
By applying findings from an investigation into English-speaking children’s
ability to comprehend verbal riddles, this paper advances an established model
of children’s humour development by expanding stage 5 to include four sub-
stages: (a) lexical and phonological ambiguities; (b) morphological ambiguities;
(c) syntactic ambiguities and (d) idiomatic ambiguities. Sixty children equally
divided from three British school Year Groups: Year 2 (aged 6-7), Year 4 (8-9)
and Year 6 (10-11) participated. Their understanding of riddles was measured (a)
receptively through a multiple-choice task in which they were required to
identify an ambiguous punchline and (b) productively through a verbal
explanation task in which they were required to explain their understanding of a
riddle containing an ambiguous word/phrase. Responses were analyzed using a
linguistically based classification system and explanations are offered as to why
some ambiguity types are easier/harder for young children to comprehend.
Keywords: child humour, verbal riddles, ambiguities, children’s cognitive and social
development, literacy development
1. Introduction
Studies on children’s humour development, especially within the field of
linguistics, have been sparse in recent decades despite the body of evidence
to suggest that humour is effective in developing creativity and divergent
thinking (Ziv, 1976, 1983, 1988), developing higher order reading skills
(Yuill, 1998; Zipke, 2007; 2009), raising reading levels (Yuill, 1998; Zipke,
2008) and developing confidence in social communication and interactions
(Graham, Papa & Brooks 1992, Nezlek & Derks 2001). Humour is also
known to be a motivational factor in literacy development, especially in the
riddle form as it is perceived as fun, enjoyable and the joking format most
favoured and recognised by young children (Wolfstein, 1954; Zipke, 2007,
2008).
The shortage of recent research on children’s verbal humour development
from within the field of linguistics is somewhat paradoxical given that verbal
humour is intrinsically based within the language through which it is
1 Giulia Baker is a researcher, teacher and former Literacy Coordinator who has taught
across all School Year Groups in the UK, as well as at undergraduate level. Her interests
lie in the ways different types of ambiguity-based humour can be used in the classroom
to motivate pupils and facilitate learning.
2 Michelle Aldridge is a Reader at Cardiff University, she lectures in psycholinguistics &
forensic linguistics. Her research focusses on the communicative experiences of
vulnerable people including children and adults with a communication disorder. She
writes widely in journals such as cognitive linguistics, forensic linguistics, health and
communication and language awareness. Corresponding author: AldridgeM@cardiff.ac.uk
Received : 20.08.2022
Accepted : 12.11.2022
Published : 30.11.2022
mailto:AldridgeM@cardiff.ac.uk
Children’s humour development Baker & Aldridge
573
communicated and arises from the manipulation of different linguistic
structures. Here, we investigate the gap in both the literature and our
understanding about the way in which different linguistic phenomena affect
children’s humour comprehension through a robust linguistically based
classification system (Baker and Aldridge 2021) and, in so doing, the
originality of this paper lies in advancing our theoretical understanding of
children’s humour development, together with progressing the final stage of
a recognized model of humour development (McGee 1979/2002).
Given our linguistic focus, we acknowledge the work achieved through
Raskin’s (1985) Semantic Script Theory of Humour (SSTH), later developed
with Attardo (1991) as the General Theory of Verbal Humour (GTVH) to
facilitate categorization of all verbally expressed jokes. The jokes (riddles) we
employ in this study are verbally expressed but are also specific in that they
rely upon the manipulation of linguistic phenomena for their humour. This
type of verbal joke must be ‘understood’ or made sense of in order that any
humorous response be activated. We therefore adopt the Incongruity
Resolution (IR) theory of humour (Suls 1972, 1983) as our framework
because it is based upon understanding, ‘the intellectual part of the humour
reaction which is an indispensable basis for amusement’ (Bariaud 1989, p.
20).
Within this framework, we adopt McGhee’s humour model which, although
formulated in 1979 and revised in 2002, still comprises the established
framework for children’s humour development. We focus upon McGhee’s
final stage which typically represents humour development around seven
years of age as we are interested in capturing children’s understanding of
verbal incongruities, multiple meanings, and ambiguities necessary for
processing verbal riddles (Bariaud, 1989; McGhee,1977, 1979, 2002; Shultz,
1974; Shultz & Horibe, 1974). To contextualize this stage, we outline, in
Table 1, McGhee’s model of humour development which comprises all five
stages through which children sequentially progress.
Table1
Summary of McGhee’s Revised Stages of Humour Development (2002)
Stage 0: Laughter without humour (0 to 6 months)
Stage 1: Laughter at the attachment figure (6 to 12-15 months)
Stage 2: Incongruous actions towards objects (12-15 months to 3-5
years)
Stage 3: Incongruous labelling of objects or actions (2 to 3-4/5
years)
Stage 4: Conceptual incongruity (3 to 5 years)
Pre riddle stage: transition period (5 to 6 -7 years)
Stage 5: Riddles and jokes (double meanings): 6-7 to 10-11 years
• Multiple meanings: ambiguity (in puns); first signs
of logic (in riddles)
Journal of Child Language Acquisition and Development – JCLAD
Vol: 10 Issue: 3 572-600, 2022
ISSN: 2148-1997
574
“Understanding” humour during the final stage (Five) of this model requires
complex processing as the listener must think about the linguistic
components of the riddle and be aware of their impact on the listener. The
development of metalinguistic skills (cf. Tunmer, Pratt & Herriman 1984)
indicates that the child is not only using language to communicate but must
also know what they are using and to what end (e.g., Hoppe & Kess 1982).
More specifically then, in these riddles participants must identify that some
linguistic unit has two (or more) meanings; that is, they must locate an
incongruity (i.e., an ambiguous word or phrase) and be able to go back and
forth between these meanings to determine how they both make sense
within a single context (i.e., they ‘resolve’ - or make sense of - the
incongruity). Dissecting a humorous text in this way, requires the listener to
work harder in terms of cognitive and linguistic processing than does a non-
ambiguous text and it is argued (e.g. Bariaud, 1989; Fowles & Glanz, 1977;
McGhee, 1971a, 1972, 1977, 1979, 2002) that children are typically seven
years of age before they start to develop the requisite cognitive skills to
process such ambiguities and thus transition from Stage Four to Stage Five
(McGhee, 1979, 2002). Associated with this ability is, of course, the
development of pragmatics and social cognition (Hyter, 2017), that is, at very
least, the listener must have a theory of mind to appreciate that the speaker
is engaging in humour, deviating from the fundamental principle that
speakers are cooperative in communication (Grice, 1989) and able to engage
in fantasy assimilation (McGhee, 1979).
In addition to evolving cognitive, social cognition and metalinguistic skills,
children’s general language development also influences humour
advancement. Specifically, in order to progress to the final humour stage,
children must understand the ways in which sounds, words, morphology
and syntax operate together to communicate meaning, knowledge which is
acquired during the pre-school years (Berger, 2011; Hoff, 2015; Kaplan,
1998) and continues throughout the primary years as they acquire later
aspects of linguistic proficiency, such as idiomatic language, (Ackerman,
1982; Cain, Towse & Knight, 2009; Gibbs, 1987; Levorato & Cacciari, 1995;
Prinz, 1983).
Although acquisition of the above skills is rapid during pre and early school
years, verbal humour is not fully understood during this period (McGhee
1979, 2002; Bariaud 1989). Whilst children may be able to understand jokes
expressed through the medium of language (‘verbally expressed’ humour),
they are not yet able to understand jokes reliant upon the manipulation of
linguistic properties (‘verbal humour’) which rely upon punning, double
entendres and ambiguities (e.g., ‘Why can’t a car play football?’ ‘Because it’s
only got one boot.’). This type of understanding only starts to develop as
they mature cognitively and linguistically and enter the final humour stage
(McGhee 1979, 2002). Even then, not all ambiguity-based humour is
understood with the same facility since ambiguity is manifested at varying
levels of language and requires different types of linguistic proficiency to be
understood. Some ambiguity types are reported to be easier to comprehend
than others (Shultz & Pilon, 1973; Shultz, 1974; Shultz & Horibe, 1974;
Hirsh-Pasek, Gleitman & Gleitman, 1978; Yuill, 1998) but earlier research
findings on humour comprehension are inconclusive, resulting in little
Children’s humour development Baker & Aldridge
575
consensus as to which types of ambiguities are most readily understood by
young children.
Part of the prevailing lack of consensus may be because previous research
has often been carried out in the field of psychology (as opposed to
linguistics) where different types of language-specific ambiguities were
included with emphasis centered upon developing a framework to account
for children’s humour development, rather than upon the properties of
language being tested. Earlier studies frequently lack ambiguity definitions,
which leads us to question which language phenomena they embody. Even
when included, definitions, as described below, are often broad and generic
and typically not interpreted or applied in the same way. This has resulted
in ‘the same linguistic phenomena being tested under different ambiguity
classifications and different linguistic phenomena being tested under the
same ambiguity classification’ (Baker & Aldridge 2021, p.239).
Taking, for example, lexical ambiguity, this ambiguity has often been used in
studies on children’s humour/ambiguity comprehension but historically
interpreted in different ways both within and across studies: Shultz and
Pilon (1973, p.728) describe it as occurring ‘when a given lexical item has
more than one semantic interpretation’, classifying ‘bank/ bank’ (river or
financial) as lexical ambiguity but place ‘pears/pairs’ in the category of
phonological ambiguity (despite their containing no sound modification when
orally represented). Similarly, Shultz and Horibe (1974) detail phonological
ambiguity as occurring ‘when a given phonological sequence can be
interpreted in more than one way’ (Shultz & Horibe, 1974, p.14) when only
one of their examples relies upon sound modification (line/lion), the other
two relying instead upon homophony (pear/pair) and mis-parsing (eighty
cups/eight teacups). Such inconsistencies have resulted in different
linguistic phenomena being tested within the same category of ambiguity
which means reported findings do not necessarily reflect outcomes for the
linguistic phenomena they purport to test e.g., Shultz and Pilon (1973)
report that children find phonological ambiguities easiest to detect even
though some of their stimuli rely upon homonymy and the shifting of word
boundaries rather than upon sound distortion and the manipulation of
phonemes. This in turn casts doubt upon the validity of findings, especially
as ambiguities manifested through manipulation of different linguistic
phenomena require different processing skills and pose differing levels of
challenge. To overcome these limitations our data are classified according to
our linguistic classification system (Baker and Aldridge 2021).
Of particular note is the tight link between the lexicon and syntax and the
categorization of riddles such as ‘Why do leopards make rubbish thieves?
Because they’re always spotted.’ Understanding this riddle requires
identification of a lexical ambiguity located in the homonym ‘spotted’ but this
is not enough per se. Although the two homophones have identical surface
representations, spotted (‘covered with spots’) and spotted (‘was seen’), they
nonetheless have two different syntactical representations (an adjective/past
tense verb form) which means this ambiguity requires grammatical
processing at a deeper level. To make sense of the ambiguity the listener
must identify not only the homophone ‘spotted’ but differences in word class
too. Zipke (2007, p.382) acknowledges this difference by restricting lexical
Journal of Child Language Acquisition and Development – JCLAD
Vol: 10 Issue: 3 572-600, 2022
ISSN: 2148-1997
576
ambiguity to cases where ‘a word has more than one meaning without a
class violation,’ but others do not - Yuill (1998), for example, classifies the
‘spotted riddle’ as comprising lexical ambiguity – rather than syntactic – and
does not accommodate the fact this ambiguity requires additional processing
skills in order it be understood.
Other - higher order - processing skills, such as the decoding of figurative
language, have also been overlooked in earlier studies, despite ambiguities
based upon this linguistic phenomenon having been included. Researchers
do not always accommodate the fact that idiomatic ambiguities require a
specific type of processing skill - that of figurative language - and instead
include stimuli based upon this ambiguity type in different types of
categories altogether, Fowles and Glanz (1997, p.446) categorise, ‘Why didn’t
the skeleton cross the road? It didn’t have the guts’ as lexical ambiguity and
Shultz and Pilon (1973, p.30) treat ‘He stepped over the line/lion’ as being
phonologically ambiguous. Neither address the fact that idiomatic ambiguity
requires linguistic knowledge which cannot be gleaned from the sum of its
individual lexical items. This is a matter which we accommodate within the
present study.
Differences in the interpretation and application of discrete ambiguity types
as outlined above, has meant that findings regarding children’s humour
comprehension are often inconsistent (Shultz & Pilon, 1973; Shultz and
Horibe, 1974; Hirsh-Pasek et al., 1978; Yuill 1998) and this has
consequently contributed to stagnation in our theoretical understanding of
the final stage of humour development. As stated, our study addresses this
matter by using Baker and Aldridge’s (2021) robust linguistically based
classification system (see appendix A), to develop both the humour model
and our understanding of the way in which children’s humour develops. We
use these definitions because they were specifically established to improve
classification of ambiguity types for practical investigation. They
accommodate the fact that verbal humour is intrinsically embedded within
the form in which it is delivered (i.e., the language through which it is
communicated) and identify the specific linguistic phenomena through
which discrete ambiguity types are manifested. Using these definitions
allows us to be sure of the specific linguistic phenomena being tested at any
given time – and of the processing skills required in order that any ambiguity
be understood.
The three aims of this investigation were: (a) to establish how children’s
comprehension of orally delivered verbal riddles differs across the ages of six
to eleven years; (b) to determine whether ambiguity type affects the facility
with which verbal riddles are understood by these children and (c) to
establish whether our data can be explained within McGhee’s model of
humour development. To this end we use verbal riddles, each reliant upon
one of five ambiguity types (lexical, phonological, morphological, syntactic,
and idiomatic) to test children’s comprehension of humour.
We hypothesis: (1) comprehension of ambiguity types will improve across
ascending Year Groups, (2) the different ways in which linguistic properties
are manipulated to create humour will affect participants’ understanding,
and (3) ambiguities requiring more complex language processing skills will
be less readily understood.
Children’s humour development Baker & Aldridge
577
2. Methodology
2.1. Participants
Participants comprised sixty children from a mainstream English medium
primary school with twenty children taking part from school Year 2 (aged six
to seven years), Year 4 (aged eight to nine years) and Year 6 (aged ten to
eleven years). These participants were selected because they spanned the
transition period between Stage Four and Five of McGhee’s (2002) humour
model when verbal ambiguities are said to be first comprehended. Thirty-five
of the participants were male and twenty-five were female. Gender
distribution was evenly spread across Year Groups. Participants were all L1
English monolingual speakers and had no known sensory, learning, or
behavioural disability.
2.2. Ethics
Consent was granted by the Headteacher of the participating school and by
parents/guardians of every participant (each of whom was provided with a
letter detailing the nature and purpose of study and how data were to be
collected, stored, and used). All data were anonymized. The University’s
ethics committee also approved the research.
Throughout the study the utmost care was taken to ensure that the
participants did not experience discomfort or find the activities stressful
either physically or psychologically. Irrespective of consent, if the researcher
felt that a child was feeling upset or uncomfortable at any stage, the study
was stopped immediately, and the child’s participation terminated in a
sensitive manner (this happened on a single occasion during the pilot study).
Most participants appeared to find the activity enjoyable and they frequently
expressed regret about it ending.
2.3. Data collection and processing
A single researcher collected the data. This allowed for consistency in the
oral delivery of the stimuli and numbers and types of prompts. The
researcher was not known to the participants. Although a class teacher
might have provided a more familiar face for administration of the study, this
was not feasible since the process was considerably time-consuming and the
teacher was occupied within the classroom. The researcher was an
experienced practicing primary teacher, used to teaching and engaging with
children across all participating Year Groups. Two additional coders were
employed to rate the participants’ responses (explanation scores)
independently after the data had been collected.
The study took place during the school day and within the school building, a
familiar environment for participants. Participants were withdrawn from
their classrooms and taken to an alternative room where they were able to
provide individual responses without being influenced by their peers and
could be recorded without interruption. It took an average of fifteen minutes
per child to complete the activity.
Verbal riddles (short question-answer jokes) (see appendix B) were used to
test participants’ ability to identify and explain different ambiguity types.
Each riddle contained a single verbal ambiguity (a word or phrase that could
Journal of Child Language Acquisition and Development – JCLAD
Vol: 10 Issue: 3 572-600, 2022
ISSN: 2148-1997
578
be interpreted in more than one way) in its punchline3 aimed at eliciting a
humorous response. Ambiguities were all either reliant upon words and
meanings, being phonetically similar to some other string not present in the
utterance or else relied upon two identically sounding utterances being
lexically analyzable in two different ways. Pre-rating of ambiguities was
piloted through classification by undergraduate Linguistics students and
only riddles that met with full consensus were included. Each riddle
contained more than a single word punchline answer, was of a similar length
and was age appropriate (non-offensive and conceptually familiar to
participants). Three riddles were used for each ambiguity type to reduce the
possibility of participants selecting a correct punchline by chance.
Processing demands and attentional abilities contributed to a limit of 15
riddles, three per ambiguity type for each participant. All riddles were
trialled in a pilot study with participants of the same age to ensure that
language/concepts were familiar to participants across Year Groups so that
these factors would not impact upon processing skills.
2.4. Data analysis
Comprehension was the focus of the investigation, and one of the
challenging issues to address was how best to determine whether
participants had understood a riddle based on ambiguous use of language. It
was important that opportunities be maximised for participants to
communicate their understanding to the researcher. Previous studies have
addressed this issue through a variety of ways including using multiple
choice exercises (Yalisove, 1978; McGhee & Panoutsopoulou, 1990; Yuill,
1998; Zipke, 2007), participants’ recall (Yalisove, 1978; Fowles & Glanz,
1977; Yuill, 1998), participants’ explanations (McGhee, 1971b; Shultz, 1974;
Shultz & Horibe, 1974; Prentice & Fathom, 1975; Fowles & Glanz, 1977;
Hirsh-Pasek et al, 1978) and graded comprehension scores (McGhee, 1971b;
Prentice & Fathom, 1975; Fowles & Glanz, 1977; Hirsh-Pasek et al., 1978).
Each of these methods was considered.
Recall was eliminated as a comprehension measure on the basis that many
young children can recall and recite short texts verbatim (in this instance
riddle punchlines consisting of three to nine words) in so called “parrot
fashion”, often without any real processing of the understanding of their
underlying meanings (Allington & Strange, 1979). This left two other main
options, multiple choice exercises and eliciting verbal explanations (which
here incorporated graded comprehension scores). Both these methods were
trialled in a pilot study to determine their efficacy as summarized in the next
section.
2.5. Pilot study
The multiple-choice task involved nine participants aged six to eleven years
listening to a verbal riddle and then selecting one of three potential
punchlines, only one of which was the correct (original) punchline. The
verbal explanation task involved the same nine participants aged six to
eleven years listening to a verbal riddle and then explaining how this
3 Ambiguities in questions are reported as being more difficult to identify (Yalisove, 1978).
Children’s humour development Baker & Aldridge
579
coupling made sense to them within a humorous context. These nine
participants did not participate in the main study.
Analyses of results from the pilot study showed that comprehension of
ambiguities was measured in both tasks but in distinct ways. The multiple-
choice task tested comprehension in terms of being able to IDENTIFY an
ambiguity while the verbal explanation task tested comprehension in terms
of being able to EXPLAIN an ambiguity. Both required metalinguistic
application but placed different demands on participants depending upon
whether comprehension was being measured receptively (identification of an
ambiguity) or productively (explanation of an ambiguity). The multiple-choice
task alleviated language production demands relating to justification of
punchline selections but did not give any insight into reasons behind
punchline selections. The verbal explanation task, on the other hand,
afforded an insight into reasons behind punchline selections, but was
potentially more difficult for participants. It was therefore decided to
combine the two tasks in the main study. Each provided a counterbalance to
potential weakness in the other and by combining the two tasks, a richer set
of data could be obtained than had only one of the tasks been carried out.
Both ambiguity identification and explanation scores were treated as being
indicative of comprehension of ambiguities in punchlines, but differences
between the two tasks, the cognitive demands they placed upon participants,
and their respective strengths and limitations were acknowledged.
2.6. Main study procedure
Participants were tested individually by the researcher. Those who had a
joke to share with the researcher (as suggested in the initial letter requesting
consent for participation) related their joke as an initial icebreaker. The
researcher then discussed what constitutes a riddle and explained that she
was interested in finding out which types of riddles were understood by
children in different school Year Groups. It was emphasized that participants
were not taking part to give a “right” or “wrong” response but rather were
sharing their ideas and understanding about individual riddles as a
representative of their particular Year Group. A warm-up activity was carried
out during which the researcher played the role of participant and chose a
punchline from a selection of three and explained how the punchline made
sense to her by identifying an ambiguous word and then explaining the two
possible interpretations. Each participant was then given a practice try
themselves before the task began.
The researcher discussed the fact that she might well use the phrase “And
anything else?” as a prompt in the context of reminding participants that
there might be an additional meaning to be explained but stressed that
participants were not to feel pressurised if they had nothing further to add.
She reassured the children that lots of people had nothing to say when given
this prompt and that was acceptable. Participants were able to ask questions
before the study began and throughout its duration. Whilst a few
participants asked for clarification regarding procedures, most questions
asked related to the technology used to record participants’ explanations.
When the researcher was sure that each participant understood the activity
through appropriately responding to the practice riddles, she read out fifteen
Journal of Child Language Acquisition and Development – JCLAD
Vol: 10 Issue: 3 572-600, 2022
ISSN: 2148-1997
580
riddles, one by one, each with three potential punchlines (see Appendix B).
Of the three potential punchlines, one was the riddle’s correct (original)
punchline containing an ambiguity upon which the riddle depended, one
was a “plausible” punchline, and one was an “irrelevant” punchline. The
“plausible” punchline treated the riddle question as if it were a bona fide
request for information rather than one intended to elicit a humorous
response and contained a logical answer to the riddle question (e.g., ‘How did
the banana know he was ill’ PLAUSIBLE PUNCHLINE = ‘He had a high
temperature’). The “irrelevant” punchline was one which neither treated the
question as a bona fide request for information, nor as a rejoinder intended
to elicit humour (e.g., ‘How did the banana know he was ill’ IRRELEVANT
PUNCHLINE = ‘He looked out of the window’.) Both original punchline
positions and ambiguity types were presented in random order although the
order was identical for each participant within each riddle multiple choice
selection. Participants were able to hear the riddle question and any of the
punchlines as many times as desired. Participants then chose the
punchlines which, when coupled with the riddles’ interrogatives, they felt
completed the riddle. Distribution of ambiguity types was spread across the
activity to address any potential fatigue effect for riddles presented later
(although subsequent analysis indicated no significant fatigue effect).
Participants were asked to indicate if they had heard any of the riddles
before. In such instances, an alternative riddle was supplied (again with a
choice of three punchlines) based on the same ambiguity type and which
met all the other criteria for inclusion4.
After choosing a punchline, participants were recorded explaining the reason
for each punchline selection. When applicable, non-leading prompts were
used e.g. “And anything else?”, “Can you explain what you mean by . . .?” to
obtain as full a response as possible. Upon completion of the two tasks
participants were thanked for their help and participation.
2.7. Scoring
Identification scores were totalled for each type of punchline selection and
for each ambiguity type correctly identified. Chi square tests and analyses of
variance were run to determine whether differences in punchline selections
by Year Group and/or ambiguity type were statistically significant.
Only explanations relating to correct punchline selections were scored since
these directly related to the three main research aims although, explanations
for non-target (i.e., plausible, and irrelevant) selections were also analyzed
qualitatively to determine reasons why participants chose answers that were
non-target non-original punchlines.
Three independent raters used comprehension criteria (cf. Comprehension
Criteria) to score transcribed explanations. Homophones were transcribed
phonetically so as not to influence raters. Morphologically ambiguous words
4 The highest number of substitute riddles were for lexical ambiguity (22). This did not
negatively impact upon comprehension since lexical ambiguities were identified most
frequently overall in the multiple-choice and scored second highest in the verbal
explanation test (cf. results) – which corresponds with findings from studies in which this
type of ambiguity has been interpreted in a similar way.
Children’s humour development Baker & Aldridge
581
were transcribed as either one or two words according to the juncture used
by participants. Words that were notably stressed were underscored.
Gestures that were made in support of, or instead of, participants’
explanations were annotated in transcriptions.
Scores ranged from 0-2 depending on whether 1, 2 or 0 interpretations of the
linguistic word/string were judged to have been communicated. When
raters could not reach a consensus, scores were awarded according to the
most frequently provided score (on no occasion did raters award three
different scores). Scores were totalled for Year Groups and for ambiguity
types. Analyses of variance were run to determine whether differences
between Year Groups and/or ambiguity types were statistically significant.
Comprehension criteria accommodated all strategies employed by
participants to communicate understanding. It did not matter which strategy
participants used as long as it showed understanding. Meanings of words
and phrases were judged to have been understood and communicated if a
participant:
1 provided a definition/explanation of a meaning e.g. “Because um when
they say seven /eɪt/ nine its um seven starts eating other numbers but
/eɪt/ is also the number after seven, so they said seven /eɪt/ nine.”
2 used context to illustrate understanding e.g. “Because /ʧiːtəz/ cheat . . .
because it means if I had snakes and ladders, I already had a go I would
roll it. I would probably just say ‘Can I have the dice for a sec?’ turn
around, get it to number six and go ‘Yay, I got a six.’”
3 manipulated juncture to highlight differences in sound(s) e.g. “Because
it’s a milkshake you shake it and then the um oh watch- um the
blender and it makes like um the milk the milk shake and the jelly
usually wobbles and he saw the milk shake so he wobbled.”
4 made contrastive reference to substituted phonemes e.g., “it’s ‘ch’
instead of ‘sh’”
5 manipulated a root word, either inflectionally or derivationally, to
exemplify meaning e.g., used “cheat”, “cheats”, “cheating”, “cheated” to
illustrate meaning of the word “cheater”: “In card games you have
/ʧiːtəz/ cheating. Somebody cheats.”
6 deliberately articulated individual words so that they varied notably
from conventional stress patterns i.e., used word stress to draw
attention to meanings: “Because it was meant to be feeling well it’s a
banana peeling.”
7 identified word clusters as containing meanings not readily analysable
from the sum of individual parts e.g., referred to idioms as ‘sayings’,
‘terms’ and ‘phrases’: “Because flying off the handle is a phrase and the
witch flies on a broom and she flies off the handle.”
8 used a deictic or iconic gesture e.g., either pointed to something or
made a movement typically associated with a word’s meaning.
9 gave an answer that indirectly showed comprehension of one or more
meanings of an ambiguous word/phrase e.g. “I chose that one because
chickens can go like /ʧiːp ʧiːp/and a /ʧiːp/ chicken would be quite
good for Christmas dinner . . . because you don’t, cos then you have
more money to spend on presents.”
Journal of Child Language Acquisition and Development – JCLAD
Vol: 10 Issue: 3 572-600, 2022
ISSN: 2148-1997
582
Once these criteria had been applied to score explanations, results were
analysed as summarised below.
3. Findings
3.1. Findings according to year group
As shown in Table 2, none of the Year Groups correctly selected the target
(i.e., original) punchline for all fifteen riddles. More specifically, no
participant in Year 2 selected all fifteen original punchlines correctly and
only two participants in Year 4 (10%) and four participants in Year 6 (20%)
chose all fifteen correct (original) punchlines. Using the binomial
distribution (p=1/3, n=300, x=147) with Original punchlines as markers of
success, the probability of Year 2 choosing 147 Original punchlines by
chance is <0.0001%. For Years 4 and 6, whose Original punchline choices
were greater, the probability is even less. We can therefore be confident that
original punchline selections were not attributable to chance.
Table 2
Punchline Selections Made by Year Groups in the Multiple-Choice Task
Year 2 Year 4 Year 6
Total number of
punchlines chosen per
category across all Year
Groups
Original
Punchline
Selected
147
(49%)
241
(80%)
265
(88%)
653
Plausible
Punchline
Selected
118
(39%)
34
(11%)
18
(6%)
170
Irrelevant
Punchline
Selected
33
(11%)
10
(3%)
10
(3%)
53
No
Punchline
Selected
2
(1%)
15
(6%)
7
(3%)
24
A chi square test run was run on all multiple-choice selections to determine
whether there was a statistically significant difference in selection choices
made by individual Year Groups. The chi square test showed there to be a
statistically significant (increased) difference as Year Groups ascended: (x2
(4) =155.4, p < 0.01). This increase was expected given participants’ general
cognitive stages of development. There was, however, a marked difference in
the way in which this increase evolved between Years 2 (aged six to seven)
and 4 (aged eight to nine) and Years 4 (aged eight to nine) and 6 (aged ten to
eleven) as illustrated in Figure 1.
Children’s humour development Baker & Aldridge
583
Figure 1. Punchline selections made by ındividual year groups in the
multiple-choice task
Although a significant difference in the ability to identify correct (original)
punchlines existed across the three Year Groups, this difference was not
equally spread. There was a greater increase in correct punchline choices
between Years 2 and 4 than between Years 4 and 6 (the differences were 94
and 24 respectively). Two separate post-hoc chi square tests were performed,
one on the difference between multiple choice selections for Years 2 and 4
(x2(2) = 81.2, p < 0.01) and one on the difference between Years 4 and 6
(x2(2) = 6.0, p > 0.05). The difference between correct punchline selections
was statistically significant for Years 2 (aged six to seven) and 4 (aged eight
to nine) (147 correct choices as opposed to 241) but not for Years 4 (8-9) and
6 (10-11) (241 correct choices as opposed to 265). The difference in correct
(original) punchline selections was statistically significant only between the
two youngest Year Groups. This would therefore suggest a period of
accelerated development in the ability to identify ambiguity-based
punchlines to occur somewhere between the ages of six to nine years and we
will return to this point in the discussion.
Only explanations relating to correct punchline selections were scored,
although explanations for non-target (i.e., plausible, and irrelevant)
explanations were also analysed qualitatively to determine reasons why
participants had chosen non-target punchlines. Using the comprehension
criteria as detailed above, but scoring individually, three independent raters
agreed with 90.6% of the explanations scored. Following discussion, they
agreed on 98.6% of scores. For the nine explanations upon which raters did
not all agree, there was always consensus between two of the three raters –
in such instances, the most frequently awarded score was recorded. These
ratings have a Fleiss’ Kappa score of 0.98 which indicates inter-rater
reliability was exceedingly strong.
Since the number of participants being scored varied from Year Group to
Year Group (depending upon how many had correctly identified the target
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Year 2 Year 4 Year 6
Original Plausible Irrelevant None Chosen
Journal of Child Language Acquisition and Development – JCLAD
Vol: 10 Issue: 3 572-600, 2022
ISSN: 2148-1997
584
punchline), a mean score was calculated for each Year Group’s verbal
explanations. ‘2’ was the highest possible score that could be obtained in the
verbal explanation task. Mean explanation scores are illustrated in Figure 2.
Figure 2. Mean explanation scores by year group
As Figure 2 shows, the overall mean scores for verbal explanations increased
as the Year Groups ascended: 0.92 (Year 2), 1.49 (Year 4), 1.63 (Year 6). An
analysis of variance was run to determine whether significant differences
existed between overall mean scores. This proved to be the case: (F (2, 57 =
14.16, p<0.01). The difference between Years 2 and 4 (0.57) was greater than
that between Years 4 and 6 (0.14), however. Hence analyses of variance were
run on Year 2 and 4 explanation scores, and Year 4 and 6 explanation
scores, to determine whether differences between consecutive Year Groups
were statistically significant. The difference (here an ascending increase) was
statistically significant between Years 2 and 4 (F (1, 57) = 26.86 p<0.01) but
not between Years 4 and 6 (F (1, 57) = 1.45, p>0.05). This mirrored the
finding from the multiple-choice task in which an accelerated rate of
development in the ability to identify ambiguity-based punchlines was
established as occurring between Years 2 and 4.
As shown in Table 3, each ambiguity type was more frequently correctly
identified as Year Groups ascended except for lexical ambiguity. Year 4
identified this type of ambiguity more frequently than Year 2 (as was the
overall trend) but also more than Year 6 (which was out of keeping with the
trend). The difference in the number of times lexical ambiguity was correctly
identified by Year 4 over Year 6 was small (not significant), however, in
relation to overall totals (55 times as opposed to 51) and did not affect the
overall order in which ambiguity types were most frequently identified
correctly.
Children’s humour development Baker & Aldridge
585
Table 3
Correct Punchline Selections by Ambiguity Type in the Multiple-Choice Task
Lexical
Phonological Morphological
Syntactic
Idiomatic
Standard
deviation
per Year
Group
Year 2
37
(62%)
24
(40%)
28
(47%)
22
(37%)
36
(60%)
6.12
Year 4
55
(92%)
53
(88%)
46
(77%)
38
(63%)
49
(82%)
5.98
Year 6
51
(85%)
54
(90%)
53
(88%)
49
(82%)
58
(97%)
3.03
Totals
143
(79%)
131
(73%)
127
(71%)
109
(61%)
143
(79%)
12.55
Lexical and idiomatic ambiguities were identified most frequently overall
(although see later discussion on misleading identification rates for idiomatic
ambiguity), followed by phonological, morphological, and syntactic
ambiguities, respectively. A two-way analysis of variance was run on Year
Groups and ambiguity types to determine whether ambiguity type had a
significant effect on correctly identified original punchlines. Although the
analysis of variance showed there to be no statistically significant effect for
ambiguity type upon correct identification rates across the three Year
Groups (F (4,8) = 3.52, p>0.05) the raw data nonetheless suggested some
type of interaction. An analysis of variance was therefore run on each
participating Year Group: Year 2 (F (4, 76) = 26.03, p<0.01), Year 4 (F (4, 76)
= 6.05, p<0.01), Year 6 (F (4, 76) = 2.21, p>0.05). There was a statistically
significant relationship between scores and ambiguity type for Years 2 and 4,
but not for Year 6. Ambiguity type thus affected comprehension significantly
(in terms of identification rates) for the two youngest Year Groups only.
These two Year Groups both identified lexical ambiguity correctly most
frequently and syntactic ambiguity least frequently. There was variation in
identification rates for the three remaining ambiguity types, but when scores
were combined for Years 2 and 4, lexical ambiguities were identified most
frequently (92 times), followed by idiomatic (85), phonological (77)
morphological (74) and syntactic (60) ambiguities, respectively. This mirrored
the overall findings for the multiple-choice task in that lexical and idiomatic
ambiguities were identified jointly most frequently overall, and first and
second most frequently respectively when scores for Years 2 and 4 were
combined.
Journal of Child Language Acquisition and Development – JCLAD
Vol: 10 Issue: 3 572-600, 2022
ISSN: 2148-1997
586
Explanations were scored 0, 1 or 2 depending on whether raters judged one,
two (or no) parts of the ambiguity to have been explained. The maximum
potential score for an explanation was 2. Mean scores (Table 4) were
calculated for each Year Group’s scored explanations for each ambiguity type
tested and accommodated the fact that only correct selections were scored
by raters.
Table 4
Mean scores by ambiguity type in the verbal explanation task
Ambiguity Type
L
e
x
ic
a
l
A
m
b
ig
u
it
y
P
h
o
n
o
lo
g
ic
a
l
A
m
b
ig
u
it
y
M
o
r
p
h
o
lo
g
ic
a
l
A
m
b
ig
u
it
y
S
y
n
t
a
c
t
ic
A
m
b
ig
u
it
y
Id
io
m
a
t
ic
A
m
b
ig
u
it
y
A
ll
A
m
b
ig
u
it
y
T
y
p
e
s
c
o
m
b
in
e
d
Year 2
1.00
1.04
0.84
0.78
0.89
0.92
(SD = 0.097)
Year 4
1.65
1.60
1.35
1.49
1.31
1.49
(SD = 0.134)
Year 6
1.70
1.75
1.62
1.53
1.54
1.63
(SD = 0.087)
As with overall explanation scores, mean scores for each ambiguity type
increased with ascending Year Group and therefore age. The largest jump in
mean scores for each ambiguity type occurred between Year 2 and Year 4. A
two-way analysis of variance was run to determine whether differences in
mean scores between Year Groups and between ambiguity types were
significant. The analysis of variance showed differences to be statistically
significant between both Year Groups (F (4,638) = 3.78, p<0.05) and
ambiguity types (F (4,638) = 3.78, p<0.05). Of the five types of ambiguity
tested, phonological ambiguities were found to score highest overall, followed
by lexical, morphological, syntactic, and idiomatic ambiguities, respectively.
That phonological and lexical ambiguities were comprehended more than
morphological and syntactic ambiguities paralleled findings from the
multiple-choice task. There was, however, a notable discrepancy in findings
for the two tasks which related solely to the comprehension of idiomatic
ambiguity. Data from the multiple-choice task showed that a high number of
participants could correctly identify original punchlines based on idiomatic
ambiguity but the subsequent analysis of scored explanations from the
verbal explanation task contradicted this finding. More specifically, analysis
of the transcriptions revealed that participants gave almost twice as many
literal meanings of idiomatic phrases as figurative meanings (105 compared
Children’s humour development Baker & Aldridge
587
to 56). This suggests that many of the correctly identified idiomatic
punchlines had been chosen based on literal meanings only and not for the
intended wordplay contained within punchlines. (It is worth noting that
when original punchlines were rejected, participants tended to opt instead
for a plausible punchline, as often accommodated in the literal interpretation
of idiomatic words or phrases). Using multiple choice together with verbal
explanations afforded a wider insight into idiomatic ambiguity
comprehension than a single method would have achieved. Using mixed
methods in this way meant that potential limitations regarding idiomatic
ambiguity comprehension were addressed whilst improving upon previous
modes of data collection for this ambiguity type.
Accommodating the above findings into the final analysis, results from both
the multiple choice and verbal explanation task corresponded as to the types
of ambiguity types comprehended most readily by participants. Lexical and
phonological ambiguities were comprehended with most facility (although in
opposing order in the two tasks) followed by morphological, syntactic, and
idiomatic ambiguities, respectively. Notably, ambiguity types that required
additional levels of processing (morphological, syntactic, and idiomatic) were
comprehended less successfully than those that did not (lexical and
phonological ambiguities).
4. Discussion
The current study aimed (a) to establish how children’s comprehension of
orally delivered verbal riddles differs across the ages of six to eleven years;
(b) to determine whether ambiguity type affects the facility with which verbal
riddles are understood by these children and (c) to establish whether our
data can be explained within McGhee’s 2002 model of humour development,
or whether modification can be suggested. These points are now discussed
in turn.
4.1. Development of comprehension of verbal riddles according to
school year
Results from both the multiple choice and verbal explanation tasks showed
that the ability to identify and explain ambiguities increased with age but
that this increase was not equally spread across Year Groups. Differences in
ambiguity identification and explanation scores were far greater between
Years 2 (aged six to seven) and 4 (aged eight to nine) than between Years 4
(aged eight to nine) and 6 (aged ten to eleven) - and statistically significant
only between Years 2 and 4. Thus, whilst there is an incremental increase in
the ability to identify and explain ambiguities across all participating Year
Groups, it was statistically significant only for those aged six to nine years.
The results therefore support our first hypothesis that comprehension of
ambiguity types will improve across ascending Year Groups. Results
additionally showed an accelerated rate of improvement to occur between
Years 2 (6-7) and Years 4 (8-9) that does not occur between Years 4 (8-9) and
Year 6 (10-11).
Journal of Child Language Acquisition and Development – JCLAD
Vol: 10 Issue: 3 572-600, 2022
ISSN: 2148-1997
588
4.2. Development of comprehension of verbal riddles according to
ambiguity type
Not all ambiguities were comprehended with the same facility and the facility
with which they were comprehended varied according to the different ways in
which linguistic properties were manipulated. This finding supports our
second hypothesis - that the different ways in which linguistic properties are
manipulated to create humour will affect participants’ understanding, as we
now evaluate.
Lexical ambiguities were identified most frequently overall in the multiple-
choice task and syntactic ambiguities the least frequently (although see
earlier discussion regarding misleading identification scores for idiomatic
ambiguity). The finding that lexical ambiguities were identified most
frequently overall and syntactic ambiguities least frequently was duplicated
when scores were combined for Year Groups 2 and 4, both of whom had the
greatest range in scores across ambiguity types and for whom findings were
significant. Lexical ambiguity, in the current context, lay within individual
lexical items and contained no grammatical class violations. This finding
corresponds with those of Shultz (1974) and Hirsh-Pasek et al. (1978) both
of whom, by virtue of the examples they provide, interpret it in the same way
as the current study and report it as scoring highest in explanation tasks.
Consistency in interpretation and application of what constitutes a lexical
ambiguity therefore shows a similar trend in findings when compared across
these studies.
After lexical ambiguity, the next highest scoring ambiguity in the verbal
explanation task was phonological ambiguity – and this ambiguity type also
comprised the second most identified type of ambiguity. This corresponds
with the findings of Shultz and Pilon (1973) and Shultz and Horibe (1974)
who also report phonological and lexical ambiguities as scoring more highly
than other types of ambiguities – it must be borne in mind, however, that
they include in the category of phonological ambiguity both homophones and
confusion about word boundaries (Shultz & Pilon 1973, p.728, Shultz &
Horibe 1974, p.14). Such linguistic phenomena are treated in the present
investigation as comprising lexical and morphological ambiguities
respectively.
Ambiguities which required processing above lexical (lexeme) and
phonological (phoneme substitution) levels – i.e., morphological, syntactic,
and idiomatic ambiguities - all required additional levels of processing and
were all more difficult for participants to identify.
Morphological ambiguities requiring the (re)processing of word structure(s)
and boundaries in addition to semantic knowledge relating to individual
lexemes and the processing of sounds, were more difficult than lexical and
phonological ambiguities for participants to identify. This parallels the
findings of Hirsh-Pasek et al. (1978) who interpret ‘morpheme boundary
ambiguity’ in a similar fashion and who also report it as being more difficult
to explain than phonological or lexical ambiguities.
In ambiguities where grammatical class violations also occurred (e.g., a noun
became a verb) participants were required to focus on the grammatical
relationships between words at phrase level as well as at lexeme level to
Children’s humour development Baker & Aldridge
589
understand the humour, resulting in this type of ambiguity being one of the
hardest to process. It was least frequently identified and scored second
lowest in the explanation task. This parallels the findings of Shultz and
Pilon (1973) and Shultz and Horibe (1974), both of whom report ‘deep
structure’ ambiguities as being difficult for young participant to explain. It
also duplicates the findings of Yuill (1998), who reports syntactic ambiguities
as being more difficult for children to identify5. It should be noted, however
that not all Yuill’s syntactic ambiguities are included in the class of syntactic
ambiguity; she classifies ‘Why do leopards make rubbish thieves? Because
they’re always spotted’ as lexical ambiguity rather than syntactical (see
earlier discussion), which means she tests this type of linguistic phenomena
in more than one category within her study.
That syntactic ambiguity was least frequently identified (other than idiomatic
ambiguity) and hardest to explain, does not correspond with Hirsh Pasek et
al.’s (1978) findings for ‘case-labelling’6 ambiguities which they report as
being easiest for children to explain (after lexical ambiguities). The examples
they provide of this ambiguity type, however, show that some of their
‘underlying structure’ ambiguities are those which would be classified as
either lexical or idiomatic ambiguities in the current study. They do not
involve a change in word class and include fixed phrases which rely upon
both figurative and literal interpretation to elicit humour.
Idiomatic ambiguity, which involved the processing of figurative meanings in
addition to literal ones, was the most difficult ambiguity type for participants
to comprehend, although this finding was not always reflected in the
multiple-choice scores. We note that the reason for this is that the
punchlines for this ambiguity type were often wrongly chosen based on their
literal meanings only. The bias towards literal meanings of idiomatic
expressions supports findings from previous studies on children’s idiom
comprehension - Ackerman (1982), Prinz (1983), Gibbs (1987), Levorato &
Cacciari (1995), Le Sourn-Bissaoui et al. (2012) all report that children have
a bias towards literal interpretations of figurative language when processing
idioms. This bias is influenced by immature cognitive and language skills
(Levorato & Cacciari, 1995) which weaken around seven years of age when
children start to acquire new skills in decoding figurative expressions.
Given the late and protracted development of non-literal processing skill, it is
unsurprising that idiomatically ambiguous original punchline selections
were made solely based on their literal meanings by so many of the younger
children in the current study who were not cognitively advanced enough to
be able to fully process figurative language. Of note is the fact that each of
the idiomatic ambiguities contained within riddles was low in transparency.
This meant that it was harder for participants to decompose the idioms when
trying to make sense of them. The meanings of the idioms ‘a piece of cake’,
5 Of the seven types of riddle punchlines Yuill opts to test, not all specifically rely on verbal
ambiguities, however. She also tests ‘pragmatic riddles’, ‘absurdity riddles’ and
‘metalinguistic riddles’. Only three of Yuill’s classifications match those used in the
current study and some labels differ (Yuill’s word-compound’ ambiguity corresponds to our
‘morphological’ ambiguity).
6 Hirsh-Pasek’s ‘case-labelling and underlying structure’ relates to differences in word class
as incorporated by our syntactic ambiguity category.
Journal of Child Language Acquisition and Development – JCLAD
Vol: 10 Issue: 3 572-600, 2022
ISSN: 2148-1997
590
‘going up the wall’ and ‘having a screw loose’ could not be determined from
their constituent parts alone and required previous exposure in order they
be understood. It is possible that low transparency may have contributed to
them being harder for participants to process than those higher in
transparency. This would correspond with the findings of Gibbs (1987),
Nippold and Rudzinski (1993), Levorato and Cacciari (1999), Nippold and
Taylor (2002) and Cain, Oakhill and Lemmon (2005). All the idiomatic
ambiguities included in the study had been trialled in the pilot study,
however, to ensure that participants of this age were familiar with their
meanings – which they were. That they were low in transparency is
therefore less likely to have been a factor in our findings than the fact that
idiomatic ambiguities (of any type) involve more complex processing skills
than those required to process other ambiguity types and as such are a later
language skill to be acquired by children, which is reflected in the outcomes.
Comparing our findings for idiomatic ambiguity with earlier studies is
difficult since this ambiguity type is not reported as having been overtly
tested. Lack of reported findings do not mean that this type of ambiguity
has not been tested previously, however. It has, although researchers have
failed to address the fact that it requires a specific type of competence and
have instead included idiomatic ambiguities in differing categories (Fowles &
Glanz in lexical ambiguity and Shultz & Pilon in phonological ambiguity).
We have, in our study, ensured this specific linguistic phenomenon – and
requisite processing skills – has been robustly accommodated within a
category of its own and this has allowed us to determine the impact that this
type of linguistic manipulation has on children’s developing ambiguity and
humour comprehension.
As our discussion shows, there are both parallels and discrepancies to be
drawn between current outcomes and reported findings from previous
studies on children’s humour and ambiguity comprehension. Ambiguity
types have not always been interpreted in the same way, however, which
means that whilst parallels may be drawn, they do not always correspond to
the type of linguistic phenomena tested. Discrepancies also occur where
parallels might otherwise have been drawn had ambiguities been interpreted
in the same way. This highlights the need for consistency in ambiguity
classifications to facilitate comparative analysis across studies which would
then enrich our understanding of the way(s) in which different language
phenomena affect children’s developing ability to comprehend ambiguities.
This issue was addressed here by using definitions which focused specially
upon inherent language phenomena embedded within the form(s) of verbal
riddles.
Outcomes from the two tasks show that the decoding of various ambiguity
types requires different processing strategies, and that these, in turn,
depend upon the type of language phenomena exploited to elicit humour.
Some ambiguity types are more complex to understand because they require
additional levels of processing. Current findings show that the less complex
the processing demand, the more readily the ambiguity is understood
(phonological and lexical ambiguities). The more complex the processing
demand, the less readily it is understood (morphological, syntactic, and
idiomatic ambiguities respectively). This supports our third hypothesis: that
Children’s humour development Baker & Aldridge
591
ambiguities requiring more complex language processing skills will be less
readily understood.
Our findings will now be assessed against McGhee’s current model of
humour development and revisions will be made accordingly.
4.3. McGhee’s Model of humour development
Although comprehension results for Year 2 were lower than for Years 4 and
6, there were, nonetheless, a considerable number of Year 2 children (aged
6-7) who could identify and explain the ambiguity in some of the riddles.
This finding therefore corresponds with McGhee’s current framework of
humour development in which children, typically around seven years of age
are said to make the transition from Stage Four to Stage Five humour and to
begin to develop the ability to comprehend ambiguity-based jokes.
Whilst our findings support McGhee’s model of humour development as just
described, it nonetheless highlights a specific aspect of the current
framework now open to challenge. Participants in Year 2, despite having
lower identification rates than the two older Year Groups, were shown to be
capable of identifying almost half of the original punchlines correctly (147
times out of a potential 300 = 49%). Even when potentially misleading
identification rates for idiomatic ambiguity (see Table 2) were removed,
participants in Year 2 were still able to correctly identify 111 punchlines.
Participants in this Year Group ranged from six to seven years. Given that
they were able to correctly identify almost half the punchlines containing
ambiguous words/phrases, these participants are well on their way to being
able to identify ambiguity-based punchlines as opposed to simply
“begin[ning]” (McGhee, 1979, p.76) to develop understanding. Even allowing
for individual variation, this proficiency is earlier than McGhee’s model
would suggest. Therefore, whilst the current data confirms the biggest
developmental leap in understanding verbal ambiguities to occur between
the ages of six and nine years, it also suggests that children might start
consolidating their understanding of ambiguity-based humour at an earlier
age than previously purported. Whether this was because the participants
were cognitively advanced, linguistically advanced, or because the link
between cognitive and humour development is not as closely bound as
previously reported is not ascertainable from the current study but certainly
merits further investigation. Future studies, for example, might build on this
finding by testing participants younger than those that participated here to
determine whether the age at which young children start to comprehend
ambiguities is earlier than presently accounted for in McGhee’s five-stage
framework of humour development. Should this be the case, then McGhee’s
model (first formulated some forty years ago) might be modified accordingly.
4.4. Developing the humour model
In addition to both supporting and challenging the humour model, outcomes
from the multiple choice and verbal explanation tasks provided an insight
into trends in verbal ambiguity comprehension not currently accommodated
or communicated by the model in its present form. As participants made the
transition to Stage Five humour, they found some types of ambiguities
easier/more difficult to comprehend. All ambiguity types required
Journal of Child Language Acquisition and Development – JCLAD
Vol: 10 Issue: 3 572-600, 2022
ISSN: 2148-1997
592
participants to assign two different meanings to a single word/string of
words in order that they be understood, but the processes involved in
retrieving and assigning the two meanings varied according to the linguistic
phenomena upon which ambiguity types were based. This, in turn, had a
bearing on the types of ambiguity that were identified and explained most
successfully. Given this finding, it is therefore proposed (Table 5) that a set
of sequential sub-stages are added to McGhee’s Stage Five humour model.
Each proposed sub-stage (5a: lexical & phonological ambiguities; 5b.
morphological ambiguities; 5c. syntactic ambiguities & 5d. idiomatic
ambiguities) interrelates with both the language phenomena embedded
within the wording of verbal riddle (i.e., the form in which it is presented)
and the different processing demands it makes of the listener.
Table 5
Revised stages of humour development
Stage 0 (Laughter without humour: 0 to 6 months) – Stage 4
(Conceptual incongruity: 3 to 5 years)
Stage 5: Riddles and jokes (double meanings/ambiguities: 6-7 to 10-11
Years)
Stage 5a: lexical and phonological ambiguities
Stage 5b: morphological ambiguities
Stage 5c: syntactic ambiguities
Stage 5d: idiomatic ambiguities
It is noted that of these four sub-stages (5a-5d), the final three stages
comprise one type of ambiguity only, whilst the first stage comprises two:
lexical and phonological ambiguities. Both were understood more than the
three other types (morphological, syntactic, and idiomatic) in the multiple
choice and explanation tasks - but in an opposing order. It therefore remains
to be determined as to whether one type is more readily understood than the
other as children first start to comprehend ambiguity-based humour. This
might thus prove an area for future focused investigation to refine the
humour model further.
5. Conclusion
Our study leads us to conclude that the facility with which verbal
ambiguities are comprehended by young children varies according to the
different ways in which linguistic phenomena have been manipulated to
elicit humour. We have accordingly applied our findings to Stage 5 of
McGhee’s humour model within which verbal ambiguities first start to be
comprehended, to accommodate children’s developmental progress within
this specific stage.
Acknowledgement
This work was supported by the Economic and Social Research Council
(grant number ES/J500197/1)
Children’s humour development Baker & Aldridge
593
References
Ackerman, B. (1982). On Comprehending Idioms: Do Children Get the
Picture? Journal of Experimental Child Psychology, 33, 439-454.
https://doi.org/10.1016/0022-0965(82)90058-3
Allington, R., & Strange, M. (1979). Remembering is Not Necessarily
Understanding in Content Areas. Reading Horizons 20 (1), 60-64.
https://scholarworks.wmich.edu/cgi/viewcontent.cgi?article=2125&co
ntext=reading_horizons
Attardo, S., & Raskin, V. (1991). Script theory revis(it)ed: joke similarity and
joke representation model. Humor: International Journal of Humor
Research 4 (3–4), 293- 348. https://doi.org/10.1515/humr.1991.4.3-
4.293
Baker, G., & Aldridge, M. (2021). Disambiguating ambiguity: Providing a
Framework for Classifying Types of Ambiguity. Linguistics and the
Human Sciences 14 (3), 237-260.
https://journal.equinoxpub.com/LHS/article/view/19339
Bariaud, F. (1989). Age Differences in Children’s Humor. Journal of Children
in Contemporary Society 21 (112), 15-45.
https://doi.org/10.1300/J274v20n01_03
Berger, K. (2011). The Developing Person Through the Life Span (eighth
edition). New York: Worth Publishers.
Cain, K., Towse, A., & Knight, R. (2009). The Development of Idiom
Comprehension: An Investigation of Semantic and Contextual
Processing Skills. Journal of Experimental Child Psychology, 102, 280-
298. https://doi.org/10.1016/j.jecp.2008.08.001
Cain, K., Oakhill, J., & Lemmon, K. (2005). The Relation between Children’s
Reading Comprehension Level and Their Comprehension of Idioms.
Journal of Experimental Child Psychology, 90, 65-87.
https://doi.org/10.1016/j.jecp.2004.09.003
Fowles, B., & Glanz, M. (1977). Competence and Talent in Verbal Riddle
Comprehension. Journal of Child Language, 4 (3), 433-452.
https://doi.org/10.1017/S0305000900001781
Gibbs, R., Jr. (1987). Linguistic Factors in Children’s Understanding of
Idioms. Journal of Child Language, 14, 569-586.
https://doi.org/10.1017/S0305000900010291
Graham, E. E., Papa, M. J., & Brooks, G. P. (1992). Functions of humor in
conversation: Conceptualization and measurement. Western Journal of
Communication, 56 (2), 161–
183. https://doi.org/10.1080/10570319209374409
Grice, P. (1989). Studies in the way of words. Cambridge, MA: Harvard
University Press.
Hirsh-Pasek, K., Gleitman, L., & Gleitman, H. (1978). What did the Brain
Say to the Mind? A Study of the Detection and Report of Ambiguity by
Young Children. In A. Sinclair, R Jarvella & W. Levelt (Eds.) The
Child’s Conception of Language (pp.97-132). SpringerVerlag.
https://doi.org/10.1007/978-3-642-67155-5_7
Hoff, E. (2015). Language Development. In M. Bornstein & M. Lamb (Eds.),
Developmental Science (Seventh edition) (pp. 443-488). New York and
London: Psychology Press.
https://doi.org/10.1515/humr.1991.4.3-4.293
https://doi.org/10.1515/humr.1991.4.3-4.293
https://doi.org/10.1016/j.jecp.2008.08.001
https://doi.org/10.1016/j.jecp.2004.09.003
https://doi.org/10.1017/S0305000900001781
https://doi.org/10.1017/S0305000900010291
Journal of Child Language Acquisition and Development – JCLAD
Vol: 10 Issue: 3 572-600, 2022
ISSN: 2148-1997
594
Hoppe, R., Kess, J. F. (1982). The Acquisition of Metalinguistic Abilities. Paper
presented at the International Conference of the AILA Commission on
Psycholinguistics, "First and Second Language Learning: Similarities
and Difference" (Milan, Italy).
Hyter, Y. D. (2017). Pragmatic assessment and intervention in children. In L.
Cummings (Ed.), Research in clinical pragmatics (pp. 493–526). Cham:
Springer. https://link.springer.com/chapter/10.1007/978-3-319-
47489-2_19
Kaplan, P. (1998). The Human Odyssey. St Paul, MN: West Publishing
Company
Le Sourn-Bissaoui, S., Caillies, S., Bernard, S., Deleau, M., & Brulé, L.
(2012). Children’s Understanding of Ambiguous Idioms and
Conversational Perspective-Taking. Journal of Experimental Child
Psychology, 112, 437-451.
https://www.researchgate.net/publication/224865848_Children's_un
derstanding_of_ambiguous_idioms_and_conversational_perspective-
taking
Levorato, C., & Cacciari, C. (1995). The Effects of Different Tasks on the
Comprehension and Production of Idioms in Children. Journal of
Experimental Child Psychology, 60, 261-283.
https://doi.org/10.1006/jecp.1995.1041
Levorato, M. & Cacciari, C. (1999). Idiom Comprehension in Children: Are
the effects of semantic analysability and context separable. Journal of
Cognitive Psychology, 11 (1), 55-66.
https://doi.org/10.1080/713752299
McGhee, P. (1971a). The Role of Operational Thinking in Children’s
Comprehension and Appreciation of Humor. Child Development, 42
(3), 733-744. https://doi.org/10.2307/1127444
McGhee, P. (1971b). Cognitive Development and Children’s Comprehension
of Humor. Child Development, 42 (1), 123-138.
https://www.jstor.org/stable/1127069
McGhee, P. (1972). On the Cognitive Origins of Incongruity Humor: Fantasy
Assimilation Versus Reality Assimilation. In P. McGhee & J. Goldstein
(Eds.), The Psychology of Humor (pp. 61-80). New York: Academic Press
McGhee, P. (1977). A Model of the Origins and Early Development of
Incongruity-Based Humor. In A. Chapman & H. Foot (Eds.), It’s a
Funny Thing Humor (pp. 27-36). Oxford: Pergamon Press
McGhee, P. (1979). Humor, its Origin and Development. San Francisco: W. H.
Freeman and Company
McGhee, P. (2002). Understanding and Promoting the Development of
Children’s Humor. Iowa: Kendall/Hunt Publishing Company
McGhee, P., & Panoutsopoulou, T. (1990). The Role of Cognitive Factors in
Children’s Metaphor and Humor Comprehension. Humor: International
Journal of Humor Research, 3 (4), 379-402.
https://doi.org/10.1515/humr.1990.3.4.379
Nezlek, J. B., & Derks, P. (2001). Use of humor as a coping mechanism,
psychological adjustment, and social interaction. Humor: International
Journal of Humor Research, 14(4), 395–413.
https://doi.org/10.1515/humr.2001.011
https://link.springer.com/chapter/10.1007/978-3-319-47489-2_19
https://link.springer.com/chapter/10.1007/978-3-319-47489-2_19
https://doi.org/10.1006/jecp.1995.1041
https://doi.org/10.2307/1127444
https://psycnet.apa.org/doi/10.1515/humr.2001.011
Children’s humour development Baker & Aldridge
595
Nippold, M. & Rudzinski, M. (1993). Familiarity and Transparency in Idiom
Explanation: A Developmental Study of Children and Adolescents.
Journal of speech and hearing Research, 36 (4), 728-737.
https://doi.org/10.1044/jshr.3604.728
Nippold, M. A., & Taylor, C. L. (2002). Judgments of idiom familiarity and
transparency: A comparison of children and adolescents. Journal of
Speech, Language, and Hearing Research, 45(2), 384–391.
https://doi.org/10.1044/1092-4388(2002/030)
Prentice, N., & Fathom, R. (1975). Joking Riddles: A Developmental Index of
Children’s Humor. Developmental Psychology, 11 (2), 210-216.
https://doi.org/10.1037/h0076455
Prinz, P. (1983). The Development of Idiomatic Meaning in Children.
Language and Speech, 26 (3), 263-272.
https://doi.org/10.1177%2F002383098302600307
Raskin, V. (1985). Semantic Mechanisms of Humor. Dordrecht-Boston-
Lancaster: D. Reidel.
Shultz, T. (1974). Development of the Appreciation of Riddles. Child
Development, 45, 100-105. https://www.jstor.org/stable/1127755
Shultz, T., & Horibe, F. (1974). Development of the Appreciation of Verbal
Jokes. Developmental Psychology, 10 (1), 13-20.
https://doi.org/10.1037/h0035549
Shultz, T. & Pilon, R. (1973). Development of the Ability to Detect Linguistic
Ambiguity. Child Development, 44, 728-733.
https://doi.org/10.2307/1127716
Suls, J. (1972) A Two-Stage Model for the Appreciation of Jokes and
Cartoons: An Information-Processing Analysis. In Goldstein, J. and
McGhee, P. (Eds). The Psychology of Humor. Theoretical Perspectives
and Empirical Issues, (pp. 81-100). New York: Academic Press
Suls, J. (1983). Cognitive Processes in Humor Appreciation. In McGhee, P.
& Goldsten, J. (Eds.) Handbook of Humor research Volume 1: Basic
Issues, (pp. 39-57). New York: Springer-Verlag.
Tunmer, W.E., Pratt, C. & Herriman, M. L. (Eds.) (1984) Metalinguistic
Awareness in Children: Theory, Research, and Implications. Springer
Series in Language and Communication, Vol. 15. Berlin: Springer-
Verlag.
Wolfstein, M. (1954). Children’s Humor: A Psychological Analysis.
Bloomington: Indiana University Press.
Yalisove, D. (1978). The Effect of Riddles Structure on Children’s
Comprehension of Riddles. Developmental Psychology, 14 (2), 173-
180. https://doi.org/10.1037/0012-1649.14.2.173
Yuill, N. (1998). Reading and Riddling: The Role of Riddle Appreciation in
Understanding and Improving Poor Text Comprehension in Children.
Current Psychology of Cognition, 17 (2), 313-342.
http://sro.sussex.ac.uk/id/eprint/13114.
Zipke, M. (2007). The Role of Metalinguistic Awareness in the Reading
Comprehension of Sixth and Seventh Graders. Reading Psychology, 28
(4), 375-396. https://doi.org/10.1080/02702710701260615
https://doi.org/10.1044/jshr.3604.728
https://doi.org/10.1044/1092-4388(2002/030)
https://content.apa.org/doi/10.1037/h0076455
https://doi.org/10.1177%2F002383098302600307
https://psycnet.apa.org/doi/10.1037/h0035549
https://doi.org/10.2307/1127716
https://psycnet.apa.org/doi/10.1037/0012-1649.14.2.173
http://sro.sussex.ac.uk/id/eprint/13114
https://doi.org/10.1080/02702710701260615
Journal of Child Language Acquisition and Development – JCLAD
Vol: 10 Issue: 3 572-600, 2022
ISSN: 2148-1997
596
Zipke, M. (2008). Teaching Metalinguistic Awareness and Reading
Comprehension with Riddles. The Reading Teacher, 62 (2), 128-137.
https://core.ac.uk/download/pdf/56325349.pdf
Zipke, M. (2009). Using Semantic Ambiguity Instruction to Improve Third
Graders’ Metalinguistic Awareness and Reading Comprehension: An
Experimental Study. Reading Research Quarterly, 44 (3), 300-321.
https://www.jstor.org/stable/25655457
Ziv, A. (1976). Facilitating Effects of Humor on Creativity. Journal of
Educational Psychology, 68 (3), 318-322.
https://doi.org/10.1037/0022-0663.68.3.318
Ziv, A. (1983). The Influence of Humorous Atmosphere on Divergent
Thinking. Contemporary Educational Psychology 8, 68-75.
https://doi.org/10.1016/0361-476X(83)90035-8
Ziv, A. (1988). Using Humor to Develop Creative Thinking. Journal of
Children in Contemporary Society, 1-2, 99-116.
https://doi.org/10.1300/J274v20n01_07
https://doi.apa.org/doi/10.1037/0022-0663.68.3.318
https://doi.org/10.1300/J274v20n01_07
Children’s humour development Baker & Aldridge
597
Appendices
Appendix A: Ambiguity definitions
Lexical Ambiguity
Lexical ambiguity occurs solely within the alternative meaning of an
individual lexical item and does not rely upon grammatical analysis at
phrase, clause or sentence level. It occurs when an individual word has more
than one meaning without any class violation. This type of ambiguity
encompasses both homonyms and polysemes since when relayed orally, both
carry the same sound but different meanings e.g. “Why are babies good at
football?” “Because they can dribble.”
Phonological Ambiguity
Phonological ambiguity occurs when the ambiguous fragment of riddle text
has two non-identical phonetic forms for the two alternative interpretations.
The modification of the phonetic form can comprise the addition, deletion, or
substitution of a phoneme. It does not involve modification of phonetic form
across word boundaries and is contained within a single lexical item e.g.
“What do whales eat for dinner?” “Fish and ships.”
Morphological Ambiguity
Morphological ambiguity occurs when there are changes in morpheme
boundaries for the two readings of the text. Other than variation in stress or
juncture, the ambiguous fragment of the riddle has identical phonetic forms
for the two alternative interpretations e.g. “Why did the jelly wobble?”
“Because it saw the milkshake/milk shake.”
Syntactic Ambiguity
Syntactic ambiguity occurs when two different underlying syntactic
structures are mapped onto a single surface structure. The two different
syntactic representations reflect different underlying grammatical relations
between lexical items. Syntactic ambiguity relies upon grammatical analysis
at whole phrase, clause, or sentence level e.g. “How was the blind carpenter
able to see?” “He picked up his hammer and saw.”
Idiomatic Ambiguity
Idiomatic ambiguity occurs when the figurative meaning of an idiom is
blended with the literal meanings of its individual lexical components e.g.
“What does Spiderman do when he’s angry?” “He goes up the wall.”
Journal of Child Language Acquisition and Development – JCLAD
Vol: 10 Issue: 3 572-600, 2022
ISSN: 2148-1997
598
Key
Original punchline
Plausible punchline
Irrelevant punchline
Appendix B: Riddle Questions with Multiple Choice Punchline
Selections
Riddle 1 (Idiomatic Ambiguity)
Why did the robot act silly?
Because he was in a daft mood
Because he had a screw loose
Because he liked apples
Riddle 2 (Syntactic Ambiguity)
Why is six afraid of seven
Because seven eats fish
Because seven ate/eight nine
Because seven is bigger
Riddle 3 (Morphological Ambiguity)
Why couldn’t the skeleton go to the ball?
He had no body/nobody to go with
He was too cold
It was past his bedtime
Riddle 4 (Lexical Ambiguity)
Why are babies good at football?
Because they can dribble
Because they kick their legs
Because they like music
Riddle 5 (Phonological Ambiguity)
How did the banana know he was ill?
He wasn’t peeling well
He had a high temperature
He looked out of the window
Riddle 6 (Idiomatic Ambiguity)
What does spiderman do when he’s angry?
He turns on the radio
He goes up the wall
He stamps his feet
Children’s humour development Baker & Aldridge
599
Riddle 7 (Morphological Ambiguity)
Why did the jelly wobble?
Because someone shook the plate
Because it saw the milk shake/milkshake
Because it was midnight
Riddle 8 (Syntactic Ambiguity)
How was the blind carpenter able to see?
He went to the circus
He picked up his hammer and saw
He put on his glasses
Riddle 9 (Lexical Ambiguity)
When is the best time to buy chickens?
When the tide is out
When they are fresh
When they’re going cheap/cheep
Riddle 10 (Phonological Ambiguity)
What do whales eat for dinner?
Fish and ships
Tasty sea creatures
Big earrings
Riddle 11 (Syntactic Ambiguity)
Why do leopards make rubbish thieves?
Because they smell of roses
Because they always get caught
Because they’re always spotted
Riddle 12 (Idiomatic Ambiguity)
Why did the schoolboy eat his homework?’
His friend said it tasted nice
His teacher said it was a piece of cake
His mum liked singing in the bath
Riddle 13 (Lexical Ambiguity)
Why can’t you ever win at cards in the jungle?
Because there are so many wild animals competing
Because it snows on the mountains
Because there are too many cheetahs/cheaters
Riddle 14 (Morphological Ambiguity)
When are roads angry?
When the birds are singing
When they are cross roads/crossroads
When you annoy them
Journal of Child Language Acquisition and Development – JCLAD
Vol: 10 Issue: 3 572-600, 2022
ISSN: 2148-1997
600
Riddle 15 (Phonological Ambiguity)
What’s a mouse’s favourite game?
Hide and squeak
Time for bed
Hunt the Cheese