The book’s core argument is that an artificial intelligence that could equal or exceed human intelligence—sometimes called artificial general intelligence (AGI)—is for mathematical reasons impossible. It offers two specific reasons for this claim: Human intelligence is a capability of a complex dynamic system—the human brain and central nervous system. Systems of this sort cannot be modelled mathematically in a way that allows them to operate inside a computer. In supporting their claim, the authors, Jobst Landgrebe and Barry Smith, marshal evidence (...) from mathematics, physics, computer science, philosophy, linguistics, and biology, setting up their book around three central questions: What are the essential marks of human intelligence? What is it that researchers try to do when they attempt to achieve "artificial intelligence" (AI)? And why, after more than 50 years, are our most common interactions with AI, for example with our bank’s computers, still so unsatisfactory? Landgrebe and Smith show how a widespread fear about AI’s potential to bring about radical changes in the nature of human beings and in the human social order is founded on an error. There is still, as they demonstrate in a final chapter, a great deal that AI can achieve which will benefit humanity. But these benefits will be achieved without the aid of systems that are more powerful than humans, which are as impossible as AI systems that are intrinsically "evil" or able to "will" a takeover of human society. (shrink)

Augustine famously claims every word is a name. Some readers take Augustine to thereby maintain a purely referentialist semantic account according to which every word is a referential expression whose meaning is its extension. Other readers think that Augustine is no referentialist and is merely claiming that every word has some meaning. In this paper, I clarify Augustine’s arguments to the effect that every word is a name and argue that ‘every word is a name’ amounts to the claim that (...) for any word, there exist tokens of that word which are autonymous nouns. Augustine takes this to be the result of universal lexical ambiguity or equivocity and I clarify how Augustine’s account of metalinguistic discourse, which is one of the most detailed to have survived from antiquity, differs from some ancient and modern theories. (shrink)

The goal of creating Artificial General Intelligence (AGI) – or in other words of creating Turing machines (modern computers) that can behave in a way that mimics human intelligence – has occupied AI researchers ever since the idea of AI was first proposed. One common theme in these discussions is the thesis that the ability of a machine to conduct convincing dialogues with human beings can serve as at least a sufficient criterion of AGI. We argue that this very ability (...) should be accepted also as a necessary condition of AGI, and we provide a description of the nature of human dialogue in particular and of human language in general against this background. We then argue that it is for mathematical reasons impossible to program a machine in such a way that it could master human dialogue behaviour in its full generality. This is (1) because there are no traditional explicitly designed mathematical models that could be used as a starting point for creating such programs; and (2) because even the sorts of automated models generated by using machine learning, which have been used successfully in areas such as machine translation, cannot be extended to cope with human dialogue. If this is so, then we can conclude that a Turing machine also cannot possess AGI, because it fails to fulfil a necessary condition thereof. At the same time, however, we acknowledge the potential of Turing machines to master dialogue behaviour in highly restricted contexts, where what is called “narrow” AI can still be of considerable utility. (shrink)

When researchers think about the interaction between language and emotion, they typically focus on descriptive emotion words. This review demonstrates that emotion can interact with language at many levels of structure, from the sound patterns of a language to its lexicon and grammar, and beyond to how it appears in conversation and discourse. Findings are considered from diverse subfields across the language sciences, including cognitive linguistics, psycholinguistics, linguistic anthropology, and conversation analysis. Taken together, it is clear that emotional expression is (...) finely tuned to language-specific structures. Future emotion research can better exploit cross-linguistic variation to unravel possible universal principles operating between language and emotion. (shrink)

'Theory' is one of the most important words in the lexicon of contemporary sociology. Yet, their ubiquity notwithstanding, it is quite unclear what sociologists mean by the words 'theory,' 'theoretical,' and 'theorize.' I argue that confusions about the meaning of 'theory' have brought about undesirable consequences, including conceptual muddles and even downright miscommunication. In this paper I tackle two questions: what does 'theory' mean in the sociological language?; and what ought 'theory' to mean in the sociological language? I proceed in (...) five stages. First, I explain why one should ask a semantic question about 'theory.' Second, I lexicographically identify seven different senses of the word, which I distinguish by means of subscripts. Third, I show some difficulties that the current lack of semantic clarity has led sociology to. Fourth, I articulate the question, 'what ought "theory" to mean?,' which I dub the 'semantic predicament', and I consider what one can learn about it from the theory literature. Fifth, I recommend a 'semantic therapy' for sociology, and advance two arguments about SP: the principle of practical reason--SP is to a large extent a political issue, which should be addressed with the help of political mechanisms; and the principle of ontological and epistemological pluralism--the solution to SP should not be too ontologically and epistemologically demanding. (shrink)

The syllogistic figures and moods can be taken to be argument schemata as can the rules of the Stoic propositional logic. Sentence schemata have been used in axiomatizations of logic only since the landmark 1927 von Neumann paper [31]. Modern philosophers know the role of schemata in explications of the semantic conception of truth through Tarski’s 1933 Convention T [42]. Mathematical logicians recognize the role of schemata in first-order number theory where Peano’s second-order Induction Axiom is approximated by Herbrand’s Induction-Axiom (...) Schema [23]. Similarly, in first-order set theory, Zermelo’s second-order Separation Axiom is approximated by Fraenkel’s first-order Separation Schema [17]. In some of several closely related senses, a schema is a complex system having multiple components one of which is a template-text or scheme-template, a syntactic string composed of one or more “blanks” and also possibly significant words and/or symbols. In accordance with a side condition the template-text of a schema is used as a “template” to specify a multitude, often infinite, of linguistic expressions such as phrases, sentences, or argument-texts, called instances of the schema. The side condition is a second component. The collection of instances may but need not be regarded as a third component. The instances are almost always considered to come from a previously identified language (whether formal or natural), which is often considered to be another component. This article reviews the often-conflicting uses of the expressions ‘schema’ and ‘scheme’ in the literature of logic. It discusses the different definitions presupposed by those uses. And it examines the ontological and epistemic presuppositions circumvented or mooted by the use of schemata, as well as the ontological and epistemic presuppositions engendered by their use. In short, this paper is an introduction to the history and philosophy of schemata. (shrink)

In order to produce a coherent narrative, speakers must identify the characters in the tale so that listeners can figure out who is doing what to whom. This paper explores whether speakers use gesture, as well as speech, for this purpose. English speakers were shown vignettes of two stories and asked to retell the stories to an experimenter. Their speech and gestures were transcribed and coded for referent identification. A gesture was considered to identify a referent if it was produced (...) in the same location as the previous gesture for that referent. We found that speakers frequently used gesture location to identify referents. Interestingly, however, they used gesture most often to identify referents that were also uniquely specified in speech. Lexical specificity in referential expressions in speech thus appears to go hand‐in‐hand with specification in referential expressions in gesture. (shrink)

Wittgenstein did not claim that the ordinary language concept ‘game’ cannot be defined: he claimed that there are multiple definitions that can be adopted for special purposes, but no single definition applicable to all games. I will defend this interpretation of Wittgenstein’s position by showing its compatibility with a pragmatic argumentative view of definitions, and how this view accounts for the diversity of disagreeing game definitions in definitional disputes.

English as a verb-medial language has a short-before-long preference, whereas Korean and Japanese as verb-final languages show a long-before-short preference. In second language research, little is known regarding how L1 processing strategies affect the ultimate attainment of target structures. Existing work has shown that native speakers of Chinese strongly prefer to utter demonstrative-classifier phrases first in subject-extracted relatives and DCLs second in object-extracted relatives. But it remains unknown whether L2 learners with typologically different language backgrounds are able to acquire native-like (...) strategies, and how they deviate from native speakers or even among themselves. Using a phrase-assembly task, we investigated advanced L2-Chinese learners whose L1s were English, Korean, and Japanese, because English lacks individual classifiers and has postnominal relative clause, whereas Korean and Japanese have individual classifiers and prenominal RCs. Results showed that the English and Korean groups deviated from the native controls’ asymmetric pattern, but the Japanese group approximated native-like performance. Furthermore, compared to the English group, the Korean and Japanese groups favored the DCL-second configuration in SRs and ORs. No differences were found between the Korean and Japanese groups. Overall, our findings suggest that L1 processing strategies play an overarching role in L2 acquisition of asymmetric positioning of DCLs in Chinese RCs. (shrink)

Malgré les critiques de toutes sortes dont elle fut et dont elle est encore l'objet, la grammaire transformationnelle est devenue en vingt ans, sous l'impulsion de Noam Chomsky, le paradigme dominant de la linguistique contemporaine. Et, bien sûr, l'incroyable rapidité de cette ascension s'est accompagnée de nombreuses crises de croissance. Linguistes, philosophes et psychologues, ainsi amenés à s'interroger sur les fondements épistémologiques de cette étrange discipline, se sont heurtés à de redoutables énigmes métathéoriques: de quel domaine factuel la linguistique est-elle (...) la science? les grammaires sont-elles empiriquement falsifiables? quel rôle l'intuition du locuteur – et singulièrement celle du linguiste – joue-t-elle dans l'élaboration et la confirmation de la théorie? quelle relation y a-t-il entre règles et lois? quel est le rapport théorique de la linguistique à la psychologie, à la sociologie, à la neurophysiologie? … Ces questions et d'autres semblables ont donné lieu ces dernières années à une étonnante profusion de livres et d'articles, où, à propos d'une discipline particulière et d'exemples précis, sont discutés, souvent en profondeur, tous les principaux problemes de l'épistémologie contemporaine. (shrink)

Languages around the world differ in terms of the number of adnominal and pronominal demonstratives they require, as well as the factors that impact on their felicitous use. Given this cross-linguistic variation in deictic demonstrative terms, and the features that determine their felicitous use, an open question is how this is accommodated within bilingual cognition and language. In particular, we were interested in the extent to which bilingual language exposure and practice might alter the way in which a bilingual is (...) using deictic demonstratives in their first language. Recent research on language attrition suggests that L2 learning selectively affects aspects of the native language, with some domains of language competence being more vulnerable than others. If demonstratives are basic, and acquired relatively early, they should be less susceptible to change and attrition. This was the hypothesis we went on to test in the current study. We tested two groups of native Spanish speakers, a control group living in Spain and an experimental group living in Norway using the Memory game paradigm. Contra to our expectations, the results indicate a significant difference between the two groups in use of deictic terms, indicative of a change in the preferred number of terms used. This suggests that deictic referential systems may change over time under pressure from bilingual language exposure. (shrink)

The subject matter of the paper is an analysis of the semantic relations between sentence negation, performative negation, and declarations in reference to utterances which speech acts theory gives the label of representatives. Apart from linguistic-semantic analyses, empirical studies have been conducted on the manner in which sentence negation and performative negation are processed. The results of Study I demonstrate that the semantic relation between sentence negation and performative negation changes depend on the type of comment (positive vs. negative), and (...) contextual factors (type of expectations towards events being commented on). As it turned out, when the situational context suggests a negative comment by the sender, participants offer similar interpretations of utterances with sentence negation and performative negation. In Study II the participants assessed the likelihood of the occurrence of the facts spoken of by a sender who uses sentence negation or performative negation. In a context suggesting positive utterances by the sender, a clear difference emerged between sentence negation and performative negation. This difference was not present in respect of negative expectations. The results achieved confirm the assumptions of the model of conversational inference regarding the influence of context on interpretation of a message. The recorded results indicate the semantic relations between declarations, sentence negation, and performative negation, which change depending on the affective significance of the message and contextual factors. (shrink)

Experimental studies on polysemy have come to contradictory conclusions on whether words with multiple senses are stored as separate or shared mental representations. The present study examined the semantic relatedness and semantic similarity of literal and non-literal (metonymic and metaphorical) senses of three word classes: nouns, verbs, and adjectives. Two methods were used: a psycholinguistic experiment and a distributional analysis of corpus data. In the experiment, participants were presented with 6–12 short phrases containing a polysemous word in literal, metonymic, or (...) metaphorical senses and were asked to classify them so that phrases with the same perceived sense were grouped together. To investigate the impact of professional background on their decisions, participants were controlled for linguistic vs. non-linguistic education. For nouns and verbs, all participants preferred to group together phrases with literal and metonymic senses, but not any other pairs of senses. For adjectives, two pairs of senses were often grouped together: literal with metonymic, and metonymic with metaphorical. Participants with a linguistic background were more accurate than participants with non-linguistic backgrounds, although both groups shared principal patterns of sense classification. For the distributional analysis of corpus data, we used a semantic vector approach to quantify the similarity of phrases with literal, metonymic, and metaphorical senses in the corpora. We found that phrases with literal and metonymic senses had the highest degree of similarity for the three word classes, and that metonymic and metaphorical senses of adjectives had the highest degree of similarity among all word classes. These findings are in line with the experimental results. Overall, the results suggest that the mental representation of a polysemous word depends on its word class. In nouns and verbs, literal and metonymic senses are stored together, while metaphorical senses are stored separately; in adjectives, metonymic senses significantly overlap with both literal and metaphorical senses. (shrink)

We have evolved to become who we are, at least in part, due to our general drive to create new things and ideas. When seeking to improve our creations, ideas, or situations, we systematically overlook opportunities to perform subtractive changes. For example, when tasked with giving feedback on an academic paper, reviewers will tend to suggest additional explanations and analyses rather than delete existing ones. Here, we show that this addition bias is systematically reflected in English language statistics along several (...) distinct dimensions. First, we show that words associated with an increase in quantity or number (e.g., add, addition, more, most) are more frequent than words associated with a decrease in quantity or number (e.g., subtract, subtraction, less, least). Second, we show that in binomial expressions, addition-related words are mentioned first, that is, add and subtract rather than subtract and add. Third, we show that the distributional semantics of verbs of change, such as to improve and to transform, overlap more with the distributional semantics of add/increase than subtract/decrease, which suggests that change verbs are implicitly biased toward addition. Fourth, addition-related words have more positive connotations than subtraction-related words. Fifth, we demonstrate that state-of-the-art large language models, such as the Generative Pre-trained Transformer (GPT-3), are also biased toward addition. We discuss the implications of our results for research on cognitive biases and decision-making. (shrink)

This study examines the event- related brain potential (ERP) of 25 Mexican monolingual Spanish-speakers when reading Spanish sentences with single entity anaphora or complex anaphora. Complex anaphora is an expression that refer to propositions, states, facts or events while, a single entity anaphora is an expression that refers back to a concrete object. Here we compare the cognitive cost in processing a single entity anaphora [éstafeminine; La renuncia (resignation)] from a complex anaphora [estoneuter; La renuncia fue aceptada (The resignation was (...) accepted)].Éstaelicited a larger positive peak at 200 ms, andestoelicited a larger frontal negativity around 400 ms. The positivity resembles the P200 component, and its amplitude is thought to represent an interaction between predictive qualities in sentence processing (i.e., graphical similarity and frequency of occurrence). Unlike parietal negativities (typical N400), frontal negativities are thought to represent the ease by which pronouns are linked with its antecedent, and how easy the information is recovered from short-term memory. Thus, the complex anaphora recruited more cognitive resources than the single entity anaphora. We also included an ungrammatical control sentence [éstemasculine; La renuncia (resignation)] to better understand the unique processes behind complex anaphoric resolution, as opposed to just general difficulty in sentence processing. In this case, event-related potentials (ERPs) elicited byéstemasculineandéstafemininewere compared. Again,éstaelicited a larger P200. However, different from the experimental condition, a left anterior negativity (LAN) effect was observed foréste; the ungrammatical condition. Altogether, the present research provides electrophysiological evidence indicating that demonstrative pronouns with different morphosyntactic features (i.e., gender) and discourse parameters (i.e., single entity or complex referent) interact during the first stage of anaphoric processing of anaphora. This stage initiated as early as 200 milliseconds after the pronoun onset and probably ends around 400 ms. (shrink)