Multimodal simile in internet memes on X responding to the 2024 U.S. presidential election

Introduction. The purpose article is to consider the representation of the image of a scientific supervisor based on Internet memes containing a multimodal metaphor. The relevance of the topic is due to the growing interest of the interdisciplinary scientific community in the study of the Internet meme as a social, linguistic, psychological phenomenon of interpersonal communication. The scientific novelty of the work is determined by the fact that until now there have been no attempts to study the representation of the academic community from the perspective of the theory of conceptual metaphor using the material of Internet memes. Methodology and sources. The methodological basis of the study was the works of linguists and cognitive scientists J. Lakoff, M. Johnson, R. Dawkins, V. Aldrich, Ch. Forsville, B. Dancygier and L. Vandelanotte. The concept of visual metaphor and multimodality is defined, the role of multimodal metaphor in Internet discourse is described, the features of the functioning of multimodal metaphor are revealed. The material for the study was Internet memes representing various aspects of interaction in the academic environment. To systematize the units, a continuous sampling method was used, along with pragmalinguistic, contextual and semantic analysis. Results and discussion. The Internet memes containing a multimodal metaphor that represents the image of a scientific supervisor and illustrates the relationship between graduate students and their mentors were found and analyzed in the paper. To select the material, a search containing the keywords “scientific supervisor”, “Phd student”, “advisor” was performed. The sources of the memes were the podcasts “The Struggling Scientists”, “The Meming Phd”, “High impact PhD memes”. By selecting keywords, 85 Internet memes were identified in which metaphorical transfers of the above images were found. Based on the data obtained, 4 metaphorical models that most clearly illustrate the relationships between the supervisor and the student were proposed. Conclusion. The study showed that multimodal metaphor is an integral part of modern Internet communication and is an effective way of influencing the addressee. Internet memes representing the image of a scientific supervisor were analyzed and systematized; 4 metaphorical mappings (“Scientific supervisor is a loving/caring parent”, “Scientific supervisor is a Jedi”, “Scientific supervisor is a monster/maniac”, “Scientific supervisor is a bad boss”) are highlighted and described. It was revealed that the considered Internet memes are characterized by the replacement of the verbal component with a non-verbal (graphic) one in the target domain or in the source domain. The studied metaphorical models reflect the most pressing problems of the academic community. The expressiveness of the metaphor used is achieved through the use of movie characters in the source domain or target domain. The metaphorical potential of Internet memes as semantic-semiotic units of Internet discourse is undoubtedly a promising area for further study.

پژوهش حاضر به بررسی شیوۀ مفهوم­ سازی خشم و عذاب الهی در قرآن کریم می­پردازد. به این ­منظور، نخست، آیه‌های دربرگیرندة مفهوم‌های مورد نظر گرد­آوری شدند و سپس از جنبة حوزه­ های مبدأ و نیز مرحله‌ها و سناریوی خشم مورد بررسی قرارگرفتند. هدف از این بررسی، یافتن تناظرهای قابلِ درک برای انسان در قرآن کریم است. مبنای این تناظر، شیوۀ مفهوم‌سازی خشم و الگوی حاکم بر آن در حالت‌های گوناگون انسانی، مدلِ کووچش (Kövecses, 1986) است. از یافته‌های این بررسی می­توان به مفهوم ­سازی خشم در قرآن کریم با استفاده از حوزه‌های مبدأ «آتش»، «بلا و خسران»، «حیوان»، «مادۀ خوراکیِ­ تلخ و گزنده»، «تاریکی و ظلمت» و «فاصله» اشاره نمود که شباهت چشم‌گیری با حوزه‌های مبدأ خشمِ انسانی دارد. این شباهت­ها با توجه به ماهیّت متفاوت خشم در خداوند و بشر، از آن ­جهت اهمیّت دارد که سازوکارهای کلام قرآنی را در انتقال پیام الهی به مخاطب بشری نشان می­دهد. به­ بیان دیگر، در راستایِ بعد هدایت­ گری و اهمیّت انتقال پیام الهی به مخاطب انسانی، خداوند از کلامی برای ارتباط با انسان بهره می‌برد که برای گونة بشری ملموس و مأنوس بوده و حقیقت‌های معنوی و واقعیت‌های انتزاعی را با بهره‌گیری از مفاهیم عینی و تجربی به او منتقل می ­نماید تا برای مخاطب خود درک‌پذیرتر باشد. همچنین، با توجه آیه‌ها و مستندات قرآنی الگوی خشم خداوند با سناریوی خشمِ انسانی تفاوت داشته و مشتمل بر سه مرحلۀ بینش و هدایت، هشدار و اخطار و در مرحلۀ آخر عقاب است.

The theoretical basis of the article is the methodology of pictorial/visual metaphor research presented in Charles Forceville's work Pictorial Metaphor in Advertising (2006) and multimodal metaphor research proposed in his book Multimodal Metaphor (2009). Both verbal and non-verbal metaphors are investigated combining interaction theory proposed by Max Black and the principles of conceptual metaphor analysis formulated in cognitive linguistics. In a metaphor, the primary and the secondary subjects are considered equal to the target and the source domains distinguished by cognitive linguists and the result of their interaction (the properties of the secondary subject (source domain) are mapped onto the primary subject (target domain)) is a conceptual metaphor. The target domain in advertising is an item or service being promoted, while the source domain is an object whose properties are attributed to the item or the service being advertised.In the discourse of advertising metaphor is realised by verbal and non-verbal forms of communications: written language, spoken language, image, music, sound, gestures. If the target and source domains in a conceptual metaphor are expressed by means of one of the indicated forms, it is treated as a monomodal metaphor, whereas if they are expressed by more than one of them, it is regarded as a multimodal metaphor. Since in the case of pictorial metaphor one of the components is expressed verbally and the other – by means of an image, it is treated as one of the varieties of multimodal metaphor.In Lithuanian printed advertising, pictorial metaphor is used to express various concepts. In the article the following examples of conceptual metaphors are analysed: JUICE IS SUN, CAR IS ANIMAL, TILE ADHESIVE IS BINDWEED, VODKA IS A NATION/PERSON. The research has revealed that in a metaphor both the source and the target domain can be expressed using pictorial and verbal means and sometimes using both of them. As a result, both verbal and pictorial means are equally important in metaphor as their interaction makes an advertisement more persuasive and effective.

이 연구는 인지언어학에 기초하여 신어 합성어의 의미 구성과 해석에 관해 살펴보는 데 목적을 두었다. 지금까지 신어에 대한 연구는 조어론적 관점이나 통사론적 관점에서 주로 어종에 따른 분류를 하거나 합성과 파생의 관점에서 이루어져 왔다. 하지만, 인지언어학의 기본 입장에 비추어 보면 언어표현의 의미는 언어 단위 자체에 내재하는 것이 아니라 언어사용자의 정신 속에서 구성된다. 신어 합성어의 의미도 구성요소의 개별 의미와 그 합으로부터 결정되지 않으며 인지 과정에서 구성되고 해석된다. 이 연구는 신어 합성어의 구성에 대한 분석이 의미 특히, 인지언어학적인 관점에서 이루어져야 할 필요가 있음을 주장하였다. 인지언어학적 관점에서 신어의 의미 해석기제로 은유와 환유, 개념적 혼성을 통해 신어 합성어를 분석하였다. 은유는 추상적인 개념 영역(목표영역(target domain))을 다른 구체적인 개념 영역(근원영역(source domain))을 통해서 이해하는 것이다. 은유에 기초한 신어는 선행어가 은유인 신어, 선행어와 후행어 간의 관계가 은유인 신어, 후행어가 은유인 신어, 전체가 은유인 신어롤 나눌 수 있었다. 환유는 매체라는 한 개념적 실체가 동일한 이상적 인지모형 내에서 목표라는 또 다른 개념적 실체에 정신적 접근을 제공하는 인지 과정으로 환유는 형식적 환유와 내용적 환유로 나눌 수 있었다. 형식적 환유는 한 형태소나 단어의 일부가 전체 단어를 대표하는 과정을 나타내는 것으로 한 단어의 어근이나 일부가 합성어의 전체를 대표하는 경우인데 그 예가 다양하게 나타난다. 내용적 환유는(내용적 환유를 기반으로 신어들은) 선행어가 환유인 신어, 후행어가 환유인 신어, 전체가 환유인 신어, 두 구성소의 관계가 환유인 신어로 나눌 수 있다. 은유나 환유만으로는 신어에 나타나는 의미 구성이나 해석이 되지 않는 것은 개념적 혼성 이론으로 설명할 수 있다. 이 연구를 통해 은유와 환유, 개념적 혼성이 신어 구성과 해석의 핵심 기제로 작동한다는 것을 볼 수 있었다.The purpose of this study is to examine newly-coined compounds’ meaning construction and interpretation based on cognitive linguistics. Previous studies on new words have often categorized word types from the perspective of word formation or have otherwise taken a syntactic view, with a focus on synthesis and derivation. In light of cognitive linguistics’ fundamental position, linguistic expressions’ meanings are not embedded in linguistic units themselves, but rather in language users’ minds. Moreover, the meanings of newly-coined compounds are not determined by their components’ individual meanings or the compounds’ conjunction; rather they are constructed and interpreted during cognitive processes, suggesting that their construction can be analyzed based on meanings, especially from the cognitive linguistic perspective, which analyzes newly-coined compounds through metaphor, metonymy, and conceptual mixture, as the mechanisms of semantic interpretation. The metaphor facilitates understanding abstract concept (target) domains through other specific concept (source) domains. Metaphor-based newly-coined words can be divided into those with metaphor antecedents, those in which the antecedent-consequent relationship is characterized by metaphor, those in which the consequents are metaphors, and those in which both the antecedents and the consequents are metaphors, respectively. The metonymy, which is a cognitive process in which the conceptual substance of media provides mental access to another of the objectives within the same ideal cognitive model, can be either a formative or a content metonymy. The former refers to the process in which a morpheme or a part of a word represents the whole word, and the root or another part of a word represents a whole compound. There are a variety of examples for this. Based on the latter, newly-coined words can be divided into those with metonymy antecedents, those with metonymy consequents; those in which the antecedent-consequent relationship is characterized by metonymy, and those in which both the antecedents and the consequents are metonymy, respectively. The conceptual mixture theory cannot account for why the meaning construction or interpretation of newly-coined words cannot be completed with only metaphor or metonymy. This study demonstrates that metaphor, metonymy, and conceptual mixture act as core mechanisms for the construction and interpretation of newly-coined words.

Multi-source domain adaptation (MSDA) in remote sensing (RS) scene classification has recently gained significant attention in the visual recognition community. It leverages multiple well-labeled source domains to train a model capable of achieving strong generalization on the target domain with little to no labeled data from the target domain. However, the distribution shifts among multiple source domains make it more challenging to align the distributions between the target domain and all source domains concurrently. Moreover, relying solely on global alignment risks losing fine-grained information for each class, especially in the task of RS scene classification. To alleviate these issues, we present a Multi-Source Subdomain Distribution Alignment Network (MSSDANet), which introduces novel network structures and loss functions for subdomain-oriented MSDA. By adopting a two-level feature extraction strategy, this model attains better global alignment between the target domain and multiple source domains, as well as alignment at the subdomain level. First, it includes a pre-trained convolutional neural network (CNN) as a common feature extractor to fully exploit the shared invariant features across one target and multiple source domains. Secondly, a dual-domain feature extractor is used after the common feature extractor, which maps the data from each pair of target and source domains to a specific dual-domain feature space and performs subdomain alignment. Finally, a dual-domain feature classifier is employed to make predictions by averaging the outputs from multiple classifiers. Accompanied by the above network, two novel loss functions are proposed to boost the classification performance. Discriminant Semantic Transfer (DST) loss is exploited to force the model to effectively extract semantic information among target and source domain samples, while Class Correlation (CC) loss is introduced to reduce the feature confusion from different classes within the target domain. It is noteworthy that our MSSDANet is developed in an unsupervised manner for domain adaptation, indicating that no label information from the target domain is required during training. Extensive experiments on four common RS image datasets demonstrate that the proposed method achieves state-of-the-art performance for cross-domain RS scene classification. Specifically, in the dual-source and three-source settings, MSSDANet outperforms the second-best algorithm in terms of overall accuracy (OA) by 2.2% and 1.6%, respectively.

Book Review. A New Insight into Theory of Conceptual Metaphor

The article focuses on the constitutive features of political internet memes representing the 47th President of the United States, Donald Trump. All the memes analysed are responses to Trump’s post on his own social media platform, Truth Social, in which he claimed that he would be the world’s best Pope. The relevance of the study is accounted for by the need to elucidate the principles of interaction between the verbal and visual modes in metaphtonymic internet memes in achieving socio-political communicative impact, as well as establishing the contribution of these modes and verbal allusion to the creation of a humorous effect as an instrument of communicative influence. The process of communication as part of socio-political processes has been analysed in terms of the methodology of multimodal cognitive linguistics. The internet meme is viewed as an interactive text, predominantly humorous in nature, circulating on the Internet as a historically established media resource providing the possibility of combining visual and verbal semiotic modes to achieve situationally determined communicative goals. The interpretation of an internet meme is possible only in the context of current socio-political events relevant for a particular society. The humorous effect of metaphtonymic memes stems from their structure containing at least two incompatible conceptual structures that are simultaneously activated in the recipient’s mind, followed by a two-stage model of resolving incongruity. The application of the cognitive metaphor and metonymy theory to multimodal communication analysis has made it possible to identify and describe visual metaphtonymy as well as visual-verbal metaphtonymy interacting with allusion. Both types of metaphtonymy are constructed on the basis of several metonymies, each activating conceptual structures of certain experiential domains, while metaphor establishes connections between these structures by projecting features from the source domain onto the target domain. The incompatibility of the activated conceptual structures creates a humorous stimulus and highlights the incongruity between Trump’s traits and the expectations of the global mass audience regarding the President of the United States. Thus, metaphtonymic memes exert socio-political influence on recipients and function as instruments for shaping public opinion.

Cross-domain recommender systems represent an emerging research topic as users generally have interactions with items from multiple domains. One goal of a cross-domain recommender system is to improve the quality of recommendations in a target domain by using user preference information from other source domains. We observe that, in many applications, users interact with items of different types e.g., artists and tags. Each recommendation problem, for example, recommending artists or recommending tags, can be seen as a different task, or, in general, a different domain. Furthermore, for such applications, explicit feedback may not be available, while implicit feedback is readily available. To handle such applications, in this paper, we propose a novel cross-domain collaborative filtering approach, based on a regularized latent factor model, to transfer knowledge between source and target domains with implicit feedback. More specifically, we identify latent user and item factors in the source domains, and transfer the user factors to the target, while controlling the amount of knowledge transferred through regularization parameters. Experimental results on six target recommendation tasks or domains from two real-world applications show the effectiveness of our approach in improving target recommendation accuracy as compared to state-of-the-art single-domain collaborative filtering approaches. Furthermore, preliminary results also suggest that our approach can handle varying percentages of user overlap between source and target domains.

Domain-transfer learning is a machine learning task to explore a source domain data set to help the learning problem in a target domain. Usually, the source domain has sufficient labeled data, while the target domain does not. In this paper, we propose a novel domain-transfer convolutional model by mapping a target domain data sample to a proxy in the source domain and applying a source domain model to the proxy for the purpose of prediction. In our framework, we firstly represent both source and target domains to feature vectors by two convolutional neural networks and then construct a proxy for each target domain sample in the source domain space. The proxy is supposed to be matched to the corresponding target domain sample convolutional representation vector well. To measure the matching quality, we proposed to maximize their squared-loss mutual information (SMI) between the proxy and target domain samples. We further develop a novel neural SMI estimator based on a parametric density ratio estimation function. Moreover, we also propose to minimize the classification error of both source domain samples and target domain proxies. The classification responses are also smoothened by manifolds of both the source domain and proxy space. By minimizing an objective function of SMI, classification error, and manifold regularization, we learn the convolutional networks of both source and target domains. In this way, the proxy of a target domain sample can be matched to the source domain data and thus benefits from the rich supervision information of the source domain. We design an iterative algorithm to update the parameters alternately and test it over benchmark data sets of abnormal behavior detection in video, Amazon product reviews sentiment analysis, etc.

<p>Most machine learning methods assume the training and test sets to be independent and have identical distributions. However, this assumption does not always hold true in practical applications. Direct training usually induces poor performance if the training and test data have distribution shifts. To address this issue, a three-part model based on using a feature extractor, a classifier, and several domain discriminators is adopted herein. This unsupervised domain adaptation model is based on multiple adversarial learning with samples of different importance. A deep neural network is used for supervised classification learning of the source domain. Numerous adversarial networks are used to constitute the domain discriminators to align each category in the source and target domains and effectively transfer knowledge from the source domain to the target domain. Triplet loss functions—classification loss, label credibility loss, and discrimination loss—are presented to further optimize the model parameters. First, the label similarity metric is designed between the target and source domains data. Second, a credibility loss function is proposed to obtain an accurate label for the unlabeled data of the target domain under training iterations. Finally, a discrimination loss is designed for multiple adversarial domain discriminators to fully utilize the unlabeled data in the target domain during training. The discrimination loss function uses predicted label probabilities as dynamic weights for the train data. The proposed method is compared with mainstream domain adaptation approaches on four public datasets: Office-31, MNIST, USPS, and SVHN. Experimental results show that the proposed method can perform well in the target domain and improve generalization performance of the model.</p> <p> </p>

On 21 May 2019, Malawi went to the polls to elect leaders in what turned out to be highly contested presidential, parliamentary and local government elections. Compared to previous elections, the 2019 tripartite elections also featured highly on social media. Focusing on the presidential election, this article analyses how these elections were depicted in popular imagination particularly in internet memes. Our contention is that the internet memes, and by extension, other popular art forms of that period, constitute a humorous, but highly significant documentation, of the key events of that election. The memes reflect hopes, aspirations and disillusionment that characterized the 2019 polls. In terms of conceptual framework, we draw on the theory of visual rhetoric alongside scholarship on internet memes as both humour and political discourse.

In an indoor positioning system (IPS), transfer learning (TL) methods are commonly used to predict the location of mobile devices under the assumption that all training instances of the target domain are given in advance. However, this assumption has been criticized for its shortcomings in dealing with the problem of signal distribution variations, especially in a dynamic indoor environment. The reasons are: collecting a sufficient number of training instances is costly, the training instances may arrive online, the feature spaces of the target and source domains may be different, and negative knowledge may be transferred in the case of a redundant source domain. In this work, we proposed an online heterogeneous transfer learning (OHetTLAL) algorithm for IPS-based RSS fingerprinting to improve the positioning performance in the target domain by fusing both source and target domain knowledge. The source domain was refined based on the target domain to avoid negative knowledge transfer. The co-occurrence measure of the feature spaces (Cmip) was used to derive the homogeneous new feature spaces, and the features with higher weight values were selected for training the classifier because they could positively affect the location prediction of the target. Thus, the objective function was minimized over the new feature spaces. Extensive experiments were conducted on two real-world scenarios of datasets, and the predictive power of the different modeling techniques were evaluated for predicting the location of a mobile device. The results have revealed that the proposed algorithm outperforms the state-of-the-art methods for fingerprint-based indoor positioning and is found robust to changing environments. Moreover, the proposed algorithm is not only resilient to fluctuating environments but also mitigates the model's overfitting problem.

Recently, domain adaptation based on deep models has been a promising way to deal with the domains with scarce labeled data, which is a critical problem for deep learning models. Domain adaptation propagates the knowledge from a source domain with rich information to the target domain. In reality, the source and target domains are mostly unbalanced in that the source domain is more resource-rich and thus has more reliable knowledge than the target domain. However, existing deep domain adaptation approaches often pre-assume the source and target domains balanced and equally, leading to a medium solution between the source and target domains, which is not optimal for the unbalanced domain adaptation. In this paper, we propose a novel Deep Asymmetric Transfer Network (DATN) to address the problem of unbalanced domain adaptation. Specifically, our model will learn a transfer function from the target domain to the source domain and meanwhile adapting the source domain classifier with more discriminative power to the target domain. By doing this, the deep model is able to adaptively put more emphasis on the resource-rich source domain. To alleviate the scarcity problem of supervised data, we further propose an unsupervised transfer method to propagate the knowledge from a lot of unsupervised data by minimizing the distribution discrepancy over the unlabeled data of two domains. The experiments on two real-world datasets demonstrate that DATN attains a substantial gain over state-of-the-art methods.

Unsupervised domain adaptation aims to transfer knowledge from labeled source domain to unlabeled target domain. Recently, multisource domain adaptation (MDA) has begun to attract attention. Its performance should go beyond simply mixing all source domains together for knowledge transfer. In this article, we propose a novel prototype-based method for MDA. Specifically, for solving the problem that the target domain has no label, we use the prototype to transfer the semantic category information from source domains to target domain. First, a feature extraction network is applied to both source and target domains to obtain the extracted features from which the domain-invariant features and domain-specific features will be disentangled. Then, based on these two kinds of features, the named inherent class prototypes and domain prototypes are estimated, respectively. Then a prototype mapping to the extracted feature space is learned in the feature reconstruction process. Thus, the class prototypes for all source and target domains can be constructed in the extracted feature space based on the previous domain prototypes and inherent class prototypes. By forcing the extracted features are close to the corresponding class prototypes for all domains, the feature extraction network is progressively adjusted. In the end, the inherent class prototypes are used as a classifier in the target domain. Our contribution is that through the inherent class prototypes and domain prototypes, the semantic category information from source domains is transformed into the target domain by constructing the corresponding class prototypes. In our method, all source and target domains are aligned twice at the feature level for better domain-invariant features and more closer features to the class prototypes, respectively. Several experiments on public data sets also prove the effectiveness of our method.

The papers in this volume deal with the issue of how corpus data relate to the questions that cognitive linguists have typically investigated with respect to conceptual mappings. The authors in this volume investigate a wide range of issues- the coherence and function of particular metaphorical models, the interaction of form and meaning, the identification of source domains of metaphorical expressions, the relationship between metaphor and discourse, the priming of metaphors, and the historical development of metaphors. The studies deal with a variety of metaphorical and metonymic source and target domains, including the source domains SPACE, ANIMALS, BODY PARTS, ORGANIZATIONS and WAR, and the target domains VERBAL ACTIVITY, ECONOMY, EMOTIONS and POLITICS. In their studies, the authors present a variety of corpus-linguistic methods for the investigation of conceptual mappings, for example, corpora annotated for semantic categories, concordances of individual source-domain items and patterns, and concordances of target-domain items. In sum, the papers in this volume show how a wide range of corpus-linguistic methods can be used to investigate a variety of issues in cognitive linguistics; the combination of corpus methods with a cognitive-linguistic view of metaphor and metonymy yields new answers to old questions (and to new questions) about the relationship between language as a conceptual phenomenon and language as a textual phenomenon.