Considering that the conveyance of mental thought through language can be modified with emotive enhancement, we can presume that language use and its emotive qualities are indeed to some capacity indicative of mental condition and emotional state, since linguistic capabilities are the product of cognitive processes that undoubtedly involve physical neurological mechanisms - on account of the multitude of neurological diseases that alter conventional health and behaviours, and typically surface from arisen physical neurological abnormalities [1] - unless the particular studied set of linguistic artifacts are preconditioned and entirely fictious, and therefore offer no true insight into the cognitive condition behind the grammatical formation, other than the presented parochial act itself. Provided that within the grammatical constituency of an authentic mental expression - the ordained mental thought that has finally been outwardly conveyed, through linguistic communication - subsists of emotional features that are contextually apparent and structurally decipherable; it is then reasonable to state, first and foremost, all cases of language use fit a distinction of emotive polarity, which is the definitive binary distinction for the ternary dimensions of feeling; secondly, a particular state of mental condition from where thoughts may originate can be a primary influence over the use of language in that specific case. Therefore, an example corpus of language and its’ corresponding emotive polarity is an indicative representation of the mental state to where the mental expression had originated from.
To attain a greater discernment of what emotive polarity is, and its importance to the investigation of this report, a better outlining of its premises would be advantageous for the future discussion of sentiment analysis. By classification of the linguistic properties in communication, the lexical elements of language usage are either emotively positive, negative, or neutral [2, 3]. If the annotation of a given lexical term is announced as positive, then the word is denoted as positively inspired, such as: happy or love. If the probabilistic interpretation of a word’s emotive features are negative, then the word corresponds to negative emotions with words like: sad or anger. If the lexical element is neither positive nor negative and is therefore classified as neutral, then the word is categorically objective and is most likely a constituent of the syntactical structure that forms part of the phrase rather than a subjective influence to the overall phrase. It is important to note that none of the word examples above are actionable or behavioural in context and are instead figuratively emotional in classification, in other words, anger is a subjective emotional word that can later denote or describe an objective actionable word like punch. It is for such reasons that different types of lexical elements, and their specific categoric classes, also including modifiers: really, require utile variations of classification methods for emotive polarity; raising the suitability of methods that involve scalar classifiers, like: extremely, as an alternative to definitive binary modes.
It is not only the aspect differences in grammatical class types that impact emotive polarity, but it is that the definitive polar annotations of emotion should not be confused with the terminological aspects of semantic negation, which can also transduce polar distinctions [4]. As the ambiguity of a positive or negative mental expression - in terms of emotion - can be obfuscated with both semantic and syntactic contributes that fulfill the contextual compositional sense of a phrase. For instance, the phrase: “I am not happy”; is in the semantical sense a true – positive – declarative statement ensuring a sense of what is being felt, and it grammatically supports the notion of being the opposite to “I am happy”, and in terms of syntactic sense, the phrase is a true – positive – negation of the phrase “I am happy”, due to the presence of the logical negative lexical element “not”. However, this discrepancy between semantic negation is not immediately recognisable to emotive polarity and therefore sentiment analysis. The negation of the word happy in the phrase modifies the original subjective emotive positivity of the word, either on its own or in a phrase. So if negation was not properly taken into account, by the implemented methods of recognising emotive polarity, especially when excluding neutral terms, this particular phrase could be probabilistically interpreted through lexical segmentation – word tokenisation – as positive due to the syntactic presence of a subjective inference: “happy”; as each lexical element in its own positional grammatic relevancy to its corresponding linguistic formation, is a part of what forms the inherent semantics that develop the mental expression, in effect, an intertwining of singular meaningful bites.
The discussed interrelating partitions of lexical constituents, that formulate the entirety of a mental expression in its grammatical form, are the compounds that entail the contextual substances that together describe the intention of the original mental thought. Effective recipiency of such linguistic features, requires a degree of comprehensibility that abides the enforced state of grammatical linguistic competency; encompassing a supposed effective acknowledgement between communicators of linguistic structures that limn the grammatical elements responsible for organising the expression of meaning in communicational forms [5]. As long as some degree of conventional grammaticality is maintained, the receiving meaning of another’s mental expression can be made sense of; however, to form a comprehensible sentence, it is not entirely based on the structural configuration of linguistic constituents alone that enable an interpreters’ meaningful recipiency. For instance, the fluent innate attributes, of someone akin to a native speaker, should be able to decipher, from a shared cultural language, the extent of an ungrammatically formed phrase from the corresponding grammatical units – words – that could be irregularly put together by a non-fluent speaker. Altogether suggesting that there is a non-categoric openness to the use of language, as a somewhat erroneous state of grammaticality does not always detract from the comprehensive semantics of a mental expression, although, on the other hand, a perfectly formed grammatical sentence can be without any meaningful sensibilities.
With regards to analysis, I will not consider the required decipherment of such intricate details to this pertinent element of linguistics until later in the writing, as it is apparent from the colloquialism of communication that language use is complex and somewhat unpredictable in its structural grammatical consistency. Consistency that is carelessly further stretched in the language demonstrated by users in environments like social media: emoticons and slang, broken word-flow with posting behaviours, and overall lack of need to abide strict grammatical forms; adding a further degree of complexity that can be better imagined than described.
It is this very admixture of formal and informal, and grammatical, and ungrammatical linguistic formations that make it computationally cumbersome to extract the true meaning of a mental expression and its intended context. Context that provides the very setting for which the conveyance of meaning corresponds to; a form of semantic evaluation that has so far only been described in terms of constant semantic features that are cultivated from the very literal manifestation of syntactic rules [6]. Rules that form acknowledged linguistic constructs that portray semantic values concerning only what is literally being communicated, for instance the earlier example from above: “I am happy”, only exhibits the declaration of an emotive state and does not pertain any significant pragmatic detail that in consequence supports context identification. Identification that enables the interpreter to gather all details intended by the mental expression so that a complete understanding of its meaning can be attained; therefore, the lack of both pragmatic detail and the existence of possible modulations to such linguistic constructs threatens the truth of the literal arguments concerning simple semantic compositions formed without sufficient pragmatic emphasis. For instance, these are examples of modulations that distort the original example phrase: “When my football team wins I am happy but today they did not”; “I am happy I beat my brother in the marathon”; “’I am happy’ is what the newly elected Minister first said to the crowd”. Such examples demonstrate the significant modification and validation that contextual additions have to the overall meaning, and effectively generate new emotive states for the above examples, in their respective order: negative, positive, and neutral. Nonetheless, example three still exemplifies the instability of semantic compositions even when there are contextual elements provided. Even though the phrase itself is objective and is not emotively inspired, as it is merely the observation of another’s emotive expression; however, in its ambiguity, there is still a possibility that the example phrase does not capture the complete subjective emphasis of the communicator and their mental bias.
[ 1 ] Zhang, F. F., Peng, W., Sweeney, J. A., Jia, Z. Y., & Gong, Q. Y. (2018) Brain structure alterations in depression: Psychoradiological evidence. CNS neuroscience & therapeutics, vol 24 (11) pp. 994–1003. DOI: 10.1111/cns.12835
[ 2 ] Paltoglou G. (2014) Sentiment Analysis in Social Media. Online Collective Action, pp. 3-17 DOI: 10.1007/978-3-7091-1340-0_1
[ 3 ] Koppel, M., & Schler, J. (2006) The importance of neutral examples for learning sentiment. Computational Intelligence, vol 22 (2) pp. 100-109. DOI: 10.1111/j.1467-8640.2006.00276.x
[ 4 ] Löbner, S. (2000) Polarity in Natural Language: Predication, Quantification and Negation in Particular and Characterizing Sentences. Linguistics and Philosophy, vol 23 pp. 213–308. DOI: 10.1023/A:1005571202592
[ 5 ] Manning, C., & Schütze, H. (1999) Foundations of Statistical Natural Language Processing. MIT Press. Cambridge, MA.
[ 6 ] Pelletier, J., & Pagin, P. (2007). Content, context and composition. Content and Context. Essays on Semantics and Pragmatics.