Chan Wing Yu, Chang Hing Yan, Cheung Pok Ho, Ho Yuk Yiu, Lau Hiu Ching
In order to fully understand how language is available to all humans despite language change and individual differences, one must first acknowledge the psychological and linguistic aspects of how humans develop language abilities.
2.1. Psychology of language
2.1.1. Evolution
The evolution of humans allows them to grasp and construct language in a more complex manner than other animals do. Its facets include biology and pedagogy.
2.1.1.1. Biology
First, humans possess specialised brain areas, namely the Broca’s area and Wernicke’s area, for language processing. The prior impacts both language comprehension and production, and may explain speech-related motor movements. Some other lesser-known functions include “language repetition, gesture production, sentence grammar and fluidity, and the interpretation of others’ actions”. Its anterior and posterior are responsible for semantics and phonology respectively (Stinnett et al., 2023). The latter is also correlated with speech production, albeit not limited to motor commands but a process known as phonological retrieval – accessing mental knowledge about phoneme sequence (Binder, 2015). Despite a similar structure in other primates, humans’ remains distinct for their unique capacity for speech production (Flinker et al., 2015). Another function exclusive to humans is the ability to think and communicate symbolically. Animals are limited to icons or indexes, whereas humans are capable of liaising with a signing system that involves sign-sign correlations where signs and their meanings are not mutually exclusive and are highly reliant on cultural contexts (Grouchy et al., 2016).
2.1.1.2. Pedagogy
Secondly, merely humans develop hierarchical syntactic structure (Zuberbühler, 2020) and pedagogical grammar rules, which make possible both comprehension and production of compound words and sentences of which their meanings vary with both composition and sequence. On the other hand, animal communication is often fixed by denotative meanings that are inherently same with interchangeable orders.
2.1.2. Neural pathways for language
Several neural pathways in humans’ nervous system work together to forge language abilities.
2.1.2.1. Arcuate fasciculus
The arcuate fasciculus is a fibre bundle composed of white matter connecting the frontal – containing Broca’s area –, parietal, and temporal – Wernicke’s area – lobes. It is primarily responsible for speech processing and its function loss may lead to conduction aphasia, a speech disorder whose affected individuals underperform in repetition and display paraphasias, despite normal comprehension and speech fluency. Its anterior end extends further into the middle frontal gyrus and the inferior precentral gyrus, which are crucial to the development of literacy and the control of voluntary motor movements (ScienceDirect, n.d.).
2.1.2.2. Superior longitudinal fasciculus
The superior longitudinal fasciculus (SLF) is another white matter bundle, binding the occipital, parietal, and temporal lobes to the frontal cortex. One of the four subcomponents, SLF III, sanctions the regulation of “somatosensory input, facial and hand fine movements, phonetics, and articulation of language”. This aids the spatial-sensory perception and memory, as well as motor skills, to produce phonemes in the appropriate place and manner of articulation (ScienceDirect, n.d.).
2.1.2.3. Inferior longitudinal fasciculus
The inferior longitudinal fasciculus (ILF) is a white-matter neural pathway that plays a key role in reading literacy and lexical and semantic processing. First, it is essential to the recognition of visual word forms, with lesions causing alexia – the inability to understand written words. Besides, it relays word-related visual input from the occipital lobe to the posterior visual word form area. Research has shown that the separation of the latter owing to degenerative neural diseases will cause postnatal alexia. Other than word recognition and information relay, its excellent structural properties are positively correlated with reading fluency and comprehension. This is consistent with the phenomenon that the rate of ILF maturation in children predicts individual differences in reading literacies. Secondly, as a pathway for the semantic ventral stream, the inferior longitudinal fasciculus enables the acquisition of new semantic mapping, explained by research findings in which ILF structural variability accounts for the success rate of word learning, as well as the extensiveness of semantic memory attained from autobiographical perception and the capacity of lexical retrieval (Herbet et al., 2018).
2.1.3. Psychological models of language processing
Multiple psychological models have been developed to theorise human language processing.
2.1.3.1. Information Processing models
The Information Processing models describe the way humans attend to, perceive, consolidate and retrieve information in a systematic mechanism. One exemplification is the “Sausage Machine” two-stage model (Frazier & Fodor, 1978). This model proposes that humans use parsing to comprehend syntactic structures, where listeners first assign auditory information to phrases and phrases to nodes using mental templates – preliminary phrase packager – and then attach higher nodes to connect them into a complete phrase marker – sentence structure supervisor.
(Generated from Padlet AI)
2.1.3.2. Information Processing models
While the Information Processing models propose the deconstruction of sentences into simpler qualities, the Interaction Activation and Competition model (McClelland & Rumelhart, 1981) conceptualises the multi-level nature of language processing, where humans employ bottom-up processing and top-down processing simultaneously. For instance, a letter within a meaningful word is perceived more quickly than that letter within a chain of letters compiled in the wrong sequence of the same word. This phenomenon results from the working memory influenced by the expectation of the word spelling from the individual’s lexical knowledge.
2.1.3.3. Connectionist models
Connectionist models emphasise that language processing is enabled through patterns of activity in interconnected neurons. These patterns resemble the structured characteristics of language. For example, for speech recognition, the connectionist TRACE model (McClelland & Elman, 1986) suggests that the first layer of neurons convert auditory input into “time-varying acoustic-phonetic representations of a word”, then activates the second layer of neurons responsible for corresponding phonemes, and finally the third layer to create a pattern of neural activity that identifies the word-level unit (Joanisse & McClelland, 2015). Another connectionist model, the Spreading Activation model hypothesises that the association of words and ideas are consolidated when activation of the neurons responsible for the original source word is spread across other neurons through neural connections. The repetitive activations of these further neurons will strengthen the transmission of the energy when the base neuron is triggered, thus allowing for semantic network once the neural conditioning is complete (American Psychological Association, n.d.).
2.2.1.1. Context
The first key difference is context. In continuous speech, the contextual information is always more detailed. Listeners can rely on surrounding words to conclude the meaning of ambiguous or unclear segments. For isolated speech segments, listeners may find it hard and challenging to understand what messages the speaker tries to tell us.
2.2.1.2. Coarticulation
The second key difference is coarticulation. It means the fact that the pronunciation of a sound is affected by the sounds before and after it. In continuous speech, coarticulation is more common than isolated speech segments, and sounds blend together smoothly. For isolated speech segments, they may lack the effects of coarticulation so it cannot lead to smoother spectral transitions from one speech segment to the other segments. So listeners will find it challenging to understand the intended message because of the less natural pronunciation of individual phonemes.
2.2.1.3. Prosody
The third key difference is prosody. It refers to the rhythm, stress, and intonation patterns of speech. For continuous speech,prosody is significant as it conveys different meanings and emotions. The pattern of rhythm and sound in the whole speech can help the audience to understand what the title is when the speaker’s voice rises and falls. For isolated speech segments, it is more difficult to interpret the intended tone or emphasis. For example, when someone said “the show is great!”, it means the show is really great or it means the opposite. Thus, prosody shows the speaker's attitude and it depends on his or her intonation.
2.2.1.4. Performance and recognition
The last key difference is performance and recognition. It is always easier for us to understand continuous speech when compared with isolated speech segments. It is because of humans’ unique brain, which is optimised to process speech as a continuous stream, and the presence of context. For isolated speech segments, voice recognition systems and voice commands, such as Siri and Alexa, may require listeners to spend more time to understand the message.
(2:58 - Time to eat Grandma.)
There is an example of how important it is to use prosody in the English language. It mentions why comma pause in speech is especially significant in our daily life.
2.2.2. How contexts effect perception of speech
How people perceive the meaning in speech depends greatly on contexts.
2.2.2.1. Semantic priming
The first key is semantic priming. Naturally, when you think of “nurse”, you will also think of “doctor”. When a prime word is shown, it basically activates a network of concepts and meanings in our brain’s memory section. Also, it is believed that priming will be increased when the targets are low frequency, especially the subjects that are with lower vocabulary cognition (Yap et al., 2009).
This picture shows how the word “yellow” can make children think of the word banana”. It is because in our brain, yellow and banana are closely bonded.
2.2.2.2. Syntactic expectation
The second key is syntactic expectations. The context in which a statement is used can have a huge impact on its own sentence structure. On the basis of the prior context, listeners usually rely on their expectations of how words and phrases could be fully organised into a complete sentence. These expectations may have an effect on how communication can be analysed and comprehended.
2.2.2.3. Pragmatic inference
The third key is pragmatic inference. Listeners always have to make conclusions about the intended meaning. This takes into account the speaker’s knowledge, intentions, and listeners expectations. For instance, if someone says, “I am melting”, we may assume that he or she feels hot; or “some of the pears are green” can be strengthened to “some but not all of them are green” where the speaker may find some of the pears are yellow.
2.2.2.4. Cultural and social factors
The last key is cultural and social factors. The social and cultural context of speech can influence perception. Different factors like the speaker’s identity, his or her accent, social rank, can all have a significant impact on how communication can generally be viewed and be understood.
2.2.3. Perception of isolated letters vs perception of words
The following section will talk about the difference between perception of isolated letters and the perception of words.
2.2.3.1. Unit of processing
The first key difference is the unit of processing. The individual symbols are generally processed at the level of isolated letters, when words are processed as meaningful units at the same time. When we perceive isolated letters, we normally concentrate on their visual characteristics such as the shape and orientation. When we perceive the words to our brain, we will naturally process them as a whole thing, detecting recognizable patterns and gaining access to their own semantic meaning.
2.2.3.2. Speed and efficiency
The second key difference is speed and efficiency. Perception of isolated words is always faster and easier, and it seems more efficient than word perception if we compare it with the perception of isolated words. This is because of the visual system’s ability to quickly understand the individual letters based on their own unique characteristics. Letters and visual words provide an intriguing method for investigating perceptual expertise. This category is too recent in our human history to be entrenched in the human genome (McCandliss et al., 2003). Reading and recognizing words will also demand more cognitive work which is necessary because it involves combining many letters to produce several meaningful units.
2.2.3.3. Contextual effects
The third key difference is contextual effects. The context in which a word is used can have an important effect on its own aesthesia and meaning. This is referred to as the context effect. Isolated letters may also have the problem of the lack of contextual information that might influence their interpretation. For example, during a basketball game, you say “the bat is flying through the sky”.We will know that the bat refers to the object used during the baseball game, interpreting the sentence as the baseball bat is flying through the sky. However, if you heard that “the bat is flying through the sky” from a documentary, you will immediately realise that it is referring to an animal bat. So, word meaning depends on the context in which it is being used.
Below is the example of the baseball bat and a bat.
2.2.4. Different ways people convert written words to speech
Here discusses how people convert written words into speech.
2.2.4.1. Phonemic decoding
The first method is phonemic decoding. Individual letters or groups of letters are mapped to their matching sounds in phonemic decoding. Beginning readers, especially primary students, often use this strategy widely as they learn to match the letters with their phonetic equivalents. Individuals can generate the corresponding speech sounds and pronounce the word by decoding each phoneme in a word. Phonetic decoding helps students to identify the vocabulary that they are not familiar with.
This video shows why decoding words is good for kids because it gives them the opportunity to recognize words in a faster way and meanwhile they will be able to figure out other words which they are reading for the first time in the future.
2.2.4.2. Whole word recognition
The second method is the whole word recognition. The main users are proficient readers, it means people who are able to quickly recognise the word and pronunciation from their minds. It is because of their formation of a mental lexicon and maybe they have met these vocabulary before so they are able and willing to remember or recognise them. Skilled readers, like people who are already familiar with whole word recognition, can always freely access word pronunciations without decoding every letter or syllable in a slower way. Whole word recognition strategies such as matching games may strengthen student’s memory when they have to match pictures with words.
2.2.3.3. Contextual cues
The third method is contextual cues. Readers can guess the pronunciation of the words they don’t know or have never met before. We can decide how to pronounce new words by the method of analysing the semantic and syntactic context. There are several cues, such as linguistics cues, situational cues, knowledge-based cues and visual cues. Individuals can use this strategy to reduce the opportunity of misunderstanding in daily life when we have to deal with the terms that have several unclear pronunciations.
2.2.3.4. Subvocalization
The last method is subvocalization. It is a method in which people silently articulate the words they are reading without making a speech. It is also defined as the internal speech while we meet a passage that we have never read, the reader has to imagine the sound and the whole scenario inside their brain of the word as it is being read (Carver, 1990). They mentally “say” the article in their heads while they read, in order to absorb the meaning of the text. This method aids individuals in maintaining focus and engaging with the subject, but it can also reduce our reading speed according to the above situation, especially when young students are trying to read a book that conveys an incomprehensible message.
2.3.1. Levels of word knowledge
How does the level of word knowledge contribute to the construction of syntax and semantics? The construction of syntax and semantics is a complex process that involves the integration of word knowledge and language rules.
2.3.1.1 Vocabulary
First of all, vocabulary acquisition has enhanced the variety of our word choices and the contexts. By incorporating a variety of words into their linguistic arsenal, individuals can employ a variety of vocabulary items that are more context-appropriate and diverse. This hence contributes to the construction of complex semantics.
2.3.1.2 Semantics
Another crucial factor constructing complex syntax is semantic understanding. As knowledge acquisition progresses, semantics develop to a point where a single word can have multiple meanings. For example “run” means the action of walking much faster. It also means to operate, like he is running a grocery shop. It also means the flow of liquid, like “the running water”. Extensive semantic understanding allows individuals to use words accurately and link different related vocabulary items.
2.3.1.3 Syntax
Sentence structure also diversifies the applications of complicated semantics. There are multifarious rules in languages we must obey as the construction of syntax goes on. People with higher education levels started regulating universal grammar rules for people to follow. For example, “I and my dog will go swimming”. It is wrong in terms of sentence structure. Instead, it should be "my dog and I will go swimming”.
2.3.1.4 Pragmatics
Pragmatic competence even gives implicit meanings other than the literal meanings of word collocations. With higher understanding and mastery of the languages, people start understanding the implications behind the words. For example, when we screw things up, there will always be a friend next to us, saying “Great job! You bloody genius!” Of course we will not take that as a compliment because he is mocking us. Therefore, the higher understanding of semantics helps us make sense of sarcasm and implications.
2.3.2 Semantic Networks
How did people organise different words into the semantic network ? People organise different words into a semantic network based on their understanding of the meanings and relationships between words. The process of organising words into a semantic network involves categorization, association, and the formation of conceptual hierarchies.
2.3.2.1 Protocol Theory
Firstly, the Protocol Theory suggests the organisation of semantic networks is based on simple word classification. A mental image will be automatically generated when a word comes to our mind. For example, when we see the word “happy”, several words would pop up in our heads, for example “emotions”, “adjective” or any other emotions that are similar to this prototype, like “sad”. This theory shows the simultaneous existence of similarities and differences even in the same category. The refine classification of wordings contributes to the modern semantic network.
2.3.2.2 Feature-based models
Feature-based models bring forth the theory that words are organised based on their shared characteristics. In the examples those words share the same characteristics like “furry”, “animals” or “adorable”. This model can capture the similarities by representing words as combinations of features.
2.3.2.3 Network models
Network models show words are represented as nodes in a network, and the connections or associations between words reflect their semantic relationships. For example, in a network model, the word “cat” might have strong connections to words like “meow”, “cute”, and “furry” due to their semantic similarity. Words that appear together in language tend to have stronger connections in a network model.
2.3.2.4 Cognitive linguistics
Cognitive linguistics suggests we learn the meanings of abstract words by having actual experiences and through the interactions with the world, like putting puzzles together to get the full image of those words. In the examples, “love” and “warmth” are complex ideas. When we feel the physical touch and beloved, these abstract ideas become virtual, allowing us to get a better grasp of other abstract and vague concepts like”passion” and “enthusiasm”.
Lonely people who have never felt loved so they hardly know the vague meaning of ‘love'
2.3.3 Internal Lexicon
People organise and differentiate the characteristics of a word in their internal lexicon through several components, including its form, meaning, syntactic category, and pronunciation.
2.3.3.1. Mental representation
Mental representation suggests the recognition of words is based on our imagination and our previous exposure of the words. For instance, when “apple” comes to our mind, our brain automatically gives us information and keywords associated with “juicy”, “ sweet” and “red” or its pronunciation (/'∞pal/).
The pictures shows us when we think of the word “apple” , different words just pop up in our heads non-stop.
2.3.3.2 Feature analysis
Feature analysis indicates when we identify similar words by differentiating their characteristics. For example, when we think of cats , words like “cute” ,”four-legged” and “mammals” pop up in our head, so we can tell the difference between a cat and a frog by comparing their different features.
The picture shows how we distinguish between cat and frog by matching and unmatching their features, e.g. their colours
2.3.3.3 Phonological differentiation
Phonological differentiation lets us simply distinguish one word from another based on their pronunciations . Considering “cat” and “cut” ,in the word “cat”, its vowel sound is /ae/, while for “cut” is /^/ .Therefore the phonological difference can differentiate words almost effortlessly.
2.3.3.4 Contextual differentiation
Contextual differentiation allows us to tell different meanings of the same word in different contexts. Some words have multiple meanings in different applications, we can tell the difference by distinguishing their actual meanings. For example, “room” means living space for sleeping. However, if it is used in “there’s no room in the elevator”, it means there is no extra space in the lift. So we can tell the difference in meanings even if it is the same word.
2.3.4 Lexical Access
Lexical access refers to the process of retrieving words from long-term memory during language production or comprehension. Various ways to get lexical access will be discussed as follows.
2.3.4.1 Automatic retrieval
First of all, automatic retrieval suggests that words can be pulled off of our mind automatically and unconsciously . For example , when we are asked what fruit that comes to our mind immediately, the word “apple” may pop up because it’s frequently used and highly associated with the subject “fruit”. The word “apple” would store in our memory and be available for the next retrieval in the correct context.
2.3.4.2 Phonological and orthographic cues
Phonological and orthographic cues propose we can predict and complete words with the assistance of the sentence pattern and our own knowledge. For instance, your friend gives you the following clue: ‘it is a mammal, it has a long neck and is four-legged. It starts with “gir” and ends with “fe”.’ We might access the word “giraffe” as it matches the above pattern. For orthographic cues, for example, if you read “it was found __ the website”, there is a missing word in the sentence. With the usage of our own knowledge, we know there’s supposed to be a preposition, which is “on”, helping us gain lexical access.
2.3.4.3 Contextual cues
The application of contextual cues contributes to the collection of the surrounding information in the contexts and can greatly assist us in finding the correct keywords and answers. For instance, in the sentence “I saw a massive, furry animal with huge claws and sharp teeth wandering around in the forest”, “massive”, “claws” , “forest” give us necessary and valuable information about the topic and narrow down our word choice. Inferring the clues, it can be an apex predator and wild animal.
2.3.4.4 Strategic search
Strategic search proposes that when we encounter prompts that we are not familiar with, automatic retrieval may not be the best option. Conscious efforts to search through mental lexicon would be required to narrow down word choices and find the best answer.
Assume that you are playing a game with your friends to find the most suitable word to the prompt – “fruit”, but you find you have trouble coming up with the most appropriate word. Therefore, you engage in conscious strategic search.
Step 1. Retrieval Strategy:
First you may come up with common fruit names like “pineapple”, “avocado”, but you soon realise they are not the best words for the prompt.
Step 2. Inhibition of irrelevant alternatives:
In the next stage, these words are said to be disposed of mentally as they are not the best answer for the prompt.
Step 3. Monitoring potential word choices:
Next, you might consider the alternative factors like word classes and phonological features. You may think of words like “pear”, “watermelon” and “papaya”. However, you soon realise they do not share the same phonological features with “fruit”.
Step 4. Refining research:
The word choices are fairly narrowed down, you decide to find a fruit name that shares the same phonological feature with the prompt. You come up with the word “starfruit” that shares the same criteria and phonological similarities with the prompt “fruit”.