Tracking the neural codes for words and phrases during semantic composition, working-memory storage, and retrieval

The ability to compose successive words into a meaningful phrase is a characteristic feature of human cognition, yet its neural mechanisms remain incompletely understood. Here, we analyze the cortical mechanisms of semantic composition using magnetoencephalography (MEG) while participants read one-word, two-word, and five-word noun phrases and compared them with a subsequent image. Decoding of MEG signals revealed three processing stages. During phrase comprehension, the representation of individual words was sustained for a variable duration depending on phrasal context. During the delay period, the word code was replaced by a working-memory code whose activation increased with semantic complexity. Finally, the speed and accuracy of retrieval depended on semantic complexity and was faster for surface than for deep semantic properties. In conclusion, we propose that the brain initially encodes phrases using factorized dimensions for successive words but later compresses them in working memory and requires a period of decompression to access them.

The neural ingredients for a language of thought are available

The classical notion of a 'language of thought' (LoT), advanced prominently by the philosopher Jerry Fodor, is an influential position in cognitive science whereby the mental representations underpinning thought are considered to be compositional and productive, enabling the construction of new complex thoughts from more primitive symbolic concepts. LoT theory has been challenged because a neural implementation has been deemed implausible. We disagree. Examples of critical computational ingredients needed for a neural implementation of a LoT have in fact been demonstrated, in particular in the hippocampal spatial navigation system of rodents. Here, we show that cell types found in spatial navigation (border cells, object cells, head-direction cells, etc.) provide key types of representation and computation required for the LoT, underscoring its neurobiological viability.

Overview of data preprocessing for machine learning applications in human microbiome research

Although metagenomic sequencing is now the preferred technique to study microbiome-host interactions, analyzing and interpreting microbiome sequencing data presents challenges primarily attributed to the statistical specificities of the data (e.g., sparse, over-dispersed, compositional, inter-variable dependency). This mini review explores preprocessing and transformation methods applied in recent human microbiome studies to address microbiome data analysis challenges. Our results indicate a limited adoption of transformation methods targeting the statistical characteristics of microbiome sequencing data. Instead, there is a prevalent usage of relative and normalization-based transformations that do not specifically account for the specific attributes of microbiome data. The information on preprocessing and transformations applied to the data before analysis was incomplete or missing in many publications, leading to reproducibility concerns, comparability issues, and questionable results. We hope this mini review will provide researchers and newcomers to the field of human microbiome research with an up-to-date point of reference for various data transformation tools and assist them in choosing the most suitable transformation method based on their research questions, objectives, and data characteristics.

How Complex Verbs Acquire Their Idiosyncratic Meanings

Complex verbs with the same preverb/prefix/particle that is both linguistically productive and analyzable can be compositional as well as non-compositional in meaning. For example, the English on has compositional spatial uses (put a hat on) but also a non-spatial "continuative" use, where its semantic contribution is consistent with multiple verbs (we played / worked / talked on despite the interruption). Comparable examples can be given with German preverbs or Russian prefixes, which are the main data analyzed in the present paper. The preverbs/prefixes/particles that encode non-compositional, construction-specific senses have been extensively studied; however, it is still far from clear how their semantic idiosyncrasies arise. Even when one can identify the contribution of the base, it is counterintuitive to assign the remaining sememes to the preverb/prefix/particle part. Therefore, on one hand, there seems to be an element without meaning, and on the other, there is a word sense that apparently comes from nowhere. In this article, I suggest analyzing compositional and non-compositional complex verbs as instantiations of two different types of constructions: one with an open slot for the preverb/prefix/particle and a fixed base verb and another with a fixed preverb/prefix/particle and an open slot for the base verb. Both experimental and corpus evidence supporting this decision is provided for Russian data. I argue that each construction implies its own meaning-processing model and that the actual choice between the two can be predicted by taking into account the discrepancy in probabilities of transition from preverb/prefix/particle to base and from base to preverb/prefix/particle.

The ABC-D of animal linguistics: are syntax and compositionality for real?

In several animal species, an alarm call (e.g. ABC notes in the Japanese tit Parus minor) can be immediately followed by a recruitment call (e.g. D notes) to yield a complex call that triggers a third behaviour, namely mobbing. This has been taken to be an argument for animal syntax and compositionality (i.e. the property by which the meaning of a complex expression depends on the meaning of its parts and the way they are put together). Several additional discoveries were made across species. First, in some cases, animals respond with mobbing to the order alarm-recruitment but not to the order recruitment-alarm. Second, animals sometimes respond similarly to functionally analogous heterospecific calls they have never heard before, and/or to artificial hybrid sequences made of conspecific and heterospecific calls in the same order, thus adding an argument for the productivity of the relevant rules. We consider the details of these arguments for animal syntax and compositionality and argue that, with one important exception (Japanese tit ABC-D sequences), they currently remain ambiguous: there are reasonable alternatives on which each call is a separate utterance and is interpreted as such ('trivial compositionality'). More generally, we propose that future studies should argue for animal syntax and compositionality by explicitly pitting the target theory against two deflationary analyses: the 'only one expression' hypothesis posits that there is no combination in the first place, for example just a simplex ABCD call; while the 'separate utterances' hypothesis posits that there are separate expressions (e.g. ABC and D), but that they form separate utterances and are neither syntactically nor semantically combined.

Gaps in the Lexicon Restrict Communication

Across languages, words carve up the world of experience in different ways. For example, English lacks an equivalent to the Chinese superordinate noun tiáowèipǐn, which is loosely translated as "ingredients used to season food while cooking." Do such differences matter? A conventional label may offer a uniquely effective way of communicating. On the other hand, lexical gaps may be easily bridged by the compositional power of language. After all, most of the ideas we want to express do not map onto simple lexical forms. We conducted a referential Director/Matcher communication task with adult speakers of Chinese and English. Directors provided a clue that Matchers used to select words from a word grid. The three target words corresponded to a superordinate term (e.g., beverages) in either Chinese or English but not both. We found that Matchers were more accurate at choosing the target words when their language lexicalized the target category. This advantage was driven entirely by the Directors' use/non-use of the intended superordinate term. The presence of a conventional superordinate had no measurable effect on speakers' within- or between-category similarity ratings. These results show that the ability to rely on a conventional term is surprisingly important despite the flexibility languages offer to communicate about non-lexicalized categories.

Generative cultural learning in children and adults: the role of compositionality and generativity in cultural evolution

Are human cultures distinctively cumulative because they are uniquely compositional? We addressed this question using a summative learning paradigm where participants saw different models build different tower elements, consisting of discrete actions and objects: stacking cubes (tower base) and linking squares (tower apex). These elements could be combined to form a tower that was optimal in terms of height and structural soundness. In addition to measuring copying fidelity, we explored whether children and adults (i) extended the knowledge demonstrated to additional tower elements and (ii) productively combined them. Results showed that children and adults copied observed demonstrations and applied them to novel exemplars. However, only adults in the imitation condition combined the two newly derived base and apex, relative to adults in a control group. Nonetheless, there were remarkable similarities between children's and adults' performance across measures. Composite measures capturing errors and overall generativity in children's and adults' performance produced few population by condition interactions. Results suggest that early in development, humans possess a suite of cognitive skills-compositionality and generativity-that transforms phylogenetically widespread social learning competencies into something that may be unique to our species, cultural learning; allowing human cultures to evolve towards greater complexity.

Concepts as plug & play devices

Research on concepts has focused on categorization. Categorization starts with a stimulus. Equally important are episodes that start with a thought. We engage in thinking to draw out new consequences from stored information, or to work out how to act. Each of the concepts out of which thought is constructed provides access to a large body of stored information. Access is not always just a matter of retrieving a stored belief (semantic memory). Often it depends on running a simulation. Simulation allows conceptual thought to draw on information in special-purpose systems, information stored in special-purpose computational dispositions and special-purpose representational structures. While the utility of simulation, prospection or imagination is widely appreciated, the role of concepts in the process is not well understood. This paper turns to cognitive and computational neuroscience for a model of how simulations enable thinkers to reach novel conclusions. Carried over to conceptual thought, the model suggests that concepts are 'plug & play' devices. The distinctive power of thought-driven simulation derives from the ability of concepts to plug into two kinds of structure at once: the combinatorial structure of a thought at one end and special-purpose structural representations at the other. This article is part of the theme issue 'Concepts in interaction: social engagement and inner experiences'.

An information theoretic score for learning hierarchical concepts

How do humans learn the regularities of their complex noisy world in a robust manner? There is ample evidence that much of this learning and development occurs in an unsupervised fashion via interactions with the environment. Both the structure of the world as well as the brain appear hierarchical in a number of ways, and structured hierarchical representations offer potential benefits for efficient learning and organization of knowledge, such as concepts (patterns) sharing parts (subpatterns), and for providing a foundation for symbolic computation and language. A major question arises: what drives the processes behind acquiring such hierarchical spatiotemporal concepts? We posit that the goal of advancing one's predictions is a major driver for learning such hierarchies and introduce an information-theoretic score that shows promise in guiding the processes, and, in particular, motivating the learner to build larger concepts. We have been exploring the challenges of building an integrated learning and developing system within the framework of prediction games, wherein concepts serve as (1) predictors, (2) targets of prediction, and (3) building blocks for future higher-level concepts. Our current implementation works on raw text: it begins at a low level, such as characters, which are the hardwired or primitive concepts, and grows its vocabulary of networked hierarchical concepts over time. Concepts are strings or n-grams in our current realization, but we hope to relax this limitation, e.g., to a larger subclass of finite automata. After an overview of the current system, we focus on the score, named CORE. CORE is based on comparing the prediction performance of the system with a simple baseline system that is limited to predicting with the primitives. CORE incorporates a tradeoff between how strongly a concept is predicted (or how well it fits its context, i.e., nearby predicted concepts) vs. how well it matches the (ground) "reality," i.e., the lowest level observations (the characters in the input episode). CORE is applicable to generative models such as probabilistic finite state machines (beyond strings). We highlight a few properties of CORE with examples. The learning is scalable and open-ended. For instance, thousands of concepts are learned after hundreds of thousands of episodes. We give examples of what is learned, and we also empirically compare with transformer neural networks and n-gram language models to situate the current implementation with respect to state-of-the-art and to further illustrate the similarities and differences with existing techniques. We touch on a variety of challenges and promising future directions in advancing the approach, in particular, the challenge of learning concepts with a more sophisticated structure.

Compositional Sequence Generation in the Entorhinal-Hippocampal System

Neurons in the medial entorhinal cortex exhibit multiple, periodically organized, firing fields which collectively appear to form an internal representation of space. Neuroimaging data suggest that this grid coding is also present in other cortical areas such as the prefrontal cortex, indicating that it may be a general principle of neural functionality in the brain. In a recent analysis through the lens of dynamical systems theory, we showed how grid coding can lead to the generation of a diversity of empirically observed sequential reactivations of hippocampal place cells corresponding to traversals of cognitive maps. Here, we extend this sequence generation model by describing how the synthesis of multiple dynamical systems can support compositional cognitive computations. To empirically validate the model, we simulate two experiments demonstrating compositionality in space or in time during sequence generation. Finally, we describe several neural network architectures supporting various types of compositionality based on grid coding and highlight connections to recent work in machine learning leveraging analogous techniques.

Curriculum learning for human compositional generalization

Generalization (or transfer) is the ability to repurpose knowledge in novel settings. It is often asserted that generalization is an important ingredient of human intelligence, but its extent, nature, and determinants have proved controversial. Here, we examine this ability with a paradigm that formalizes the transfer learning problem as one of recomposing existing functions to solve unseen problems. We find that people can generalize compositionally in ways that are elusive for standard neural networks and that human generalization benefits from training regimes in which items are axis aligned and temporally correlated. We describe a neural network model based around a Hebbian gating process that can capture how human generalization benefits from different training curricula. We additionally find that adult humans tend to learn composable functions asynchronously, exhibiting discontinuities in learning that resemble those seen in child development.

Statistical challenges in longitudinal microbiome data analysis

The microbiome is a complex and dynamic community of microorganisms that co-exist interdependently within an ecosystem, and interact with its host or environment. Longitudinal studies can capture temporal variation within the microbiome to gain mechanistic insights into microbial systems; however, current statistical methods are limited due to the complex and inherent features of the data. We have identified three analytical objectives in longitudinal microbial studies: (1) differential abundance over time and between sample groups, demographic factors or clinical variables of interest; (2) clustering of microorganisms evolving concomitantly across time and (3) network modelling to identify temporal relationships between microorganisms. This review explores the strengths and limitations of current methods to fulfill these objectives, compares different methods in simulation and case studies for objectives (1) and (2), and highlights opportunities for further methodological developments. R tutorials are provided to reproduce the analyses conducted in this review.

Effects of Structure and Meaning on Cortical Tracking of Linguistic Units in Naturalistic Speech

Recent research has established that cortical activity "tracks" the presentation rate of syntactic phrases in continuous speech, even though phrases are abstract units that do not have direct correlates in the acoustic signal. We investigated whether cortical tracking of phrase structures is modulated by the extent to which these structures compositionally determine meaning. To this end, we recorded electroencephalography (EEG) of 38 native speakers who listened to naturally spoken Dutch stimuli in different conditions, which parametrically modulated the degree to which syntactic structure and lexical semantics determine sentence meaning. Tracking was quantified through mutual information between the EEG data and either the speech envelopes or abstract annotations of syntax, all of which were filtered in the frequency band corresponding to the presentation rate of phrases (1.1-2.1 Hz). Overall, these mutual information analyses showed stronger tracking of phrases in regular sentences than in stimuli whose lexical-syntactic content is reduced, but no consistent differences in tracking between sentences and stimuli that contain a combination of syntactic structure and lexical content. While there were no effects of compositional meaning on the degree of phrase-structure tracking, analyses of event-related potentials elicited by sentence-final words did reveal meaning-induced differences between conditions. Our findings suggest that cortical tracking of structure in sentences indexes the internal generation of this structure, a process that is modulated by the properties of its input, but not by the compositional interpretation of its output.

Emergence of a compositional neural code for written words: Recycling of a convolutional neural network for reading

The visual word form area (VWFA) is a region of human inferotemporal cortex that emerges at a fixed location in the occipitotemporal cortex during reading acquisition and systematically responds to written words in literate individuals. According to the neuronal recycling hypothesis, this region arises through the repurposing, for letter recognition, of a subpart of the ventral visual pathway initially involved in face and object recognition. Furthermore, according to the biased connectivity hypothesis, its reproducible localization is due to preexisting connections from this subregion to areas involved in spoken-language processing. Here, we evaluate those hypotheses in an explicit computational model. We trained a deep convolutional neural network of the ventral visual pathway, first to categorize pictures and then to recognize written words invariantly for case, font, and size. We show that the model can account for many properties of the VWFA, particularly when a subset of units possesses a biased connectivity to word output units. The network develops a sparse, invariant representation of written words, based on a restricted set of reading-selective units. Their activation mimics several properties of the VWFA, and their lesioning causes a reading-specific deficit. The model predicts that, in literate brains, written words are encoded by a compositional neural code with neurons tuned either to individual letters and their ordinal position relative to word start or word ending or to pairs of letters (bigrams).

Indeterminate and Enriched Propositions in Context Linger: Evidence From an Eye-Tracking False Memory Paradigm

A sentence such as We finished the paper is indeterminate with regards to what we finished doing with the paper. Indeterminate sentences constitute a test case for two major issues regarding language comprehension: (1) how we compose sentence meaning; and (2) what is retained in memory about what we read in context over time. In an eye-tracking experiment, participants read short stories that were unexpectedly followed by one of three recognition probes: (a) an indeterminate sentence (Lisa began the book), that is identical to the one in the story; (b) an enriched but false probe (Lisa began reading the book); and (c) a contextually unrelated probe (Lisa began writing the book). The probes were presented either at the offset of the original indeterminate sentence in context or following additional neutral discourse. We measured accuracy, probe recognition time, and reading times of the probe sentences. Results showed that, at the immediate time point, participants correctly accepted the identical probes with high accuracy and short recognition times, but that this effect reversed to chance-level accuracy and significantly longer recognition times at the delayed time point. We also found that participants falsely accept the enriched probe at both time points 50% of the time. There were no reading-time differences between identical and enriched probes, suggesting that enrichment might not be an early, mandatory process for indeterminate sentences. Overall, results suggest that while context produces an enriched proposition, an unenriched proposition true to the indeterminate sentence also lingers in memory.

Tissue schematics map the specialization of immune tissue motifs and their appropriation by tumors

A schematic of a biological system, i.e., a representation of its pieces, how they are combined, and what they do, would facilitate understanding its essential organization and alteration in pathogenesis or evolution. We present a computational approach for constructing tissue schematics (TSs) from high-parameter imaging data and a biological model for interpreting them. TSs map the spatial assembly of cellular neighborhoods into tissue motifs, whose modular composition, we propose, enables the generation of complex outputs. We developed our approach in human lymphoid tissue (HLT), identifying the follicular outer zone as a potential relay between neighboring zones and a core lymphoid assembly with modifications characteristic of each HLT type. Applying the TS approach to the tumor microenvironment in human colorectal cancer identified a higher-order motif, whose mutated assembly was negatively associated with patient survival. TSs may therefore elucidate how immune architectures can be specialized and become vulnerable to reprogramming by tumors.

Communicating Compositional Patterns

How do people perceive and communicate structure? We investigate this question by letting participants play a communication game, where one player describes a pattern, and another player redraws it based on the description alone. We use this paradigm to compare two models of pattern description, one compositional (complex structures built out of simpler ones) and one noncompositional. We find that compositional patterns are communicated more effectively than noncompositional patterns, that a compositional model of pattern description predicts which patterns are harder to describe, and that this model can be used to evaluate participants’ drawings, producing humanlike quality ratings. Our results suggest that natural language can tap into a compositionally structured pattern description language.

Network Analysis of Human Brain Connectivity Reveals Neural Fingerprints of a Compositionality Bias in Signaling Systems

Compositionality is a hallmark of human language and other symbolic systems: a finite set of meaningful elements can be systematically combined to convey an open-ended array of ideas. Compositionality is not uniformly distributed over expressions in a language or over individuals' communicative behavior: at both levels, variation is observed. Here, we investigate the neural bases of interindividual variability by probing the relationship between intrinsic characteristics of brain networks and compositional behavior. We first collected functional resting-state and diffusion magnetic resonance imaging data from a large participant sample (N = 51). Subsequently, participants took part in two signaling games. They were instructed to learn and reproduce an auditory symbolic system of signals (tone sequences) associated with affective meanings (human faces expressing emotions). Signal-meaning mappings were artificial and had to be learned via repeated signaling interactions. We identified a temporoparietal network in which connection length was related to the degree of compositionality introduced in a signaling system by each player. Graph-theoretic analysis of resting-state functional connectivity revealed that, within that network, compositional behavior was associated with integration measures in 2 semantic hubs: the left posterior cingulate cortex and the left angular gyrus. Our findings link individual variability in compositional biases to variation in the anatomy of semantic networks and in the functional topology of their constituent units.

Compositional RL Agents That Follow Language Commands in Temporal Logic

We demonstrate how a reinforcement learning agent can use compositional recurrent neural networks to learn to carry out commands specified in linear temporal logic (LTL). Our approach takes as input an LTL formula, structures a deep network according to the parse of the formula, and determines satisfying actions. This compositional structure of the network enables zero-shot generalization to significantly more complex unseen formulas. We demonstrate this ability in multiple problem domains with both discrete and continuous state-action spaces. In a symbolic domain, the agent finds a sequence of letters that satisfy a specification. In a Minecraft-like environment, the agent finds a sequence of actions that conform to a formula. In the Fetch environment, the robot finds a sequence of arm configurations that move blocks on a table to fulfill the commands. While most prior work can learn to execute one formula reliably, we develop a novel form of multi-task learning for RL agents that allows them to learn from a diverse set of tasks and generalize to a new set of diverse tasks without any additional training. The compositional structures presented here are not specific to LTL, thus opening the path to RL agents that perform zero-shot generalization in other compositional domains.

Compositional Processing Emerges in Neural Networks Solving Math Problems

A longstanding question in cognitive science concerns the learning mechanisms underlying compositionality in human cognition. Humans can infer the structured relationships (e.g., grammatical rules) implicit in their sensory observations (e.g., auditory speech), and use this knowledge to guide the composition of simpler meanings into complex wholes. Recent progress in artificial neural networks has shown that when large models are trained on enough linguistic data, grammatical structure emerges in their representations. We extend this work to the domain of mathematical reasoning, where it is possible to formulate precise hypotheses about how meanings (e.g., the quantities corresponding to numerals) should be composed according to structured rules (e.g., order of operations). Our work shows that neural networks are not only able to infer something about the structured relationships implicit in their training data, but can also deploy this knowledge to guide the composition of individual meanings into composite wholes.

Default Inheritance in Modified Statements: Bias or Inference?

It is a fact that human subjects rate sentences about typical properties such as "Ravens are black" as very likely to be true. In comparison, modified sentences such as "Feathered ravens are black" receive lower ratings, especially if the modifier is atypical for the noun, as in "Jungle ravens are black". This is called the modifier effect. However, the likelihood of the unmodified statement influences the perceived likelihood of the modified statement: the higher the rated likelihood of the unmodified sentence, the higher the rated likelihood of the modified one. That means the modifier effect does not fully block default inheritance of typical properties from nouns to modified nouns. This paper discusses this inheritance effect. In particular, I ask whether it is the direct result of composing concepts from nouns, that is, a bias toward "black" when processing "raven". I report a series of experiments in which I find no evidence for a direct inheritance from composition. This supports the view that default inheritance is rather an inference than a bias.

The Perception of Relations

The world contains not only objects and features (red apples, glass bowls, wooden tables), but also relations holding between them (apples contained in bowls, bowls supported by tables). Representations of these relations are often developmentally precocious and linguistically privileged; but how does the mind extract them in the first place? Although relations themselves cast no light onto our eyes, a growing body of work suggests that even very sophisticated relations display key signatures of automatic visual processing. Across physical, eventive, and social domains, relations such as support, fit, cause, chase, and even socially interact are extracted rapidly, are impossible to ignore, and influence other perceptual processes. Sophisticated and structured relations are not only judged and understood, but also seen - revealing surprisingly rich content in visual perception itself.

Context Breeds False Memories for Indeterminate Sentences

What are the roles of semantic and pragmatic processes in the interpretation of sentences in context? And how do we attain such interpretations when sentences are deemed indeterminate? Consider a sentence such as "Lisa began the book" which does not overtly express the activity that Lisa began doing with the book. Although it is believed that individuals compute a specified event to enrich the sentential representation - yielding, e.g., "began [reading] the book" - there is no evidence that a default event meaning is attained. Moreover, if indeterminate sentences are enriched, it is not clear where the information required to generate enriched interpretations come from. Experiment 1 showed that, in isolation, there is no default interpretation for indeterminate sentences. The experiment also showed that biasing contexts constrain event interpretations and improve plausibility judgments, suggesting that event representations for indeterminate sentences are generated by context. In Experiment 2, participants heard biasing discourse contexts and later falsely recognized foil sentences containing the biased events ("Lisa began reading the book") at the same proportion and with the same confidence as the original indeterminate sentence ("Lisa began the book"). We suggest that indeterminate sentences trigger event-enriching inferences but only in sufficiently constraining contexts. We also suggest that indeterminate sentences create two memory traces, one for the proposition consistent with the denotational, compositional meaning, and another for the proposition that is enriched pragmatically over time.

A Compositional Model of Consciousness Based on Consciousness-Only

Scientific studies of consciousness rely on objects whose existence is assumed to be independent of any consciousness. On the contrary, we assume consciousness to be fundamental, and that one of the main features of consciousness is characterized as being other-dependent. We set up a framework which naturally subsumes this feature by defining a compact closed category where morphisms represent conscious processes. These morphisms are a composition of a set of generators, each being specified by their relations with other generators, and therefore co-dependent. The framework is general enough and fits well into a compositional model of consciousness. Interestingly, we also show how our proposal may become a step towards avoiding the hard problem of consciousness, and thereby address the combination problem of conscious experiences.

Symbolic, Distributed, and Distributional Representations for Natural Language Processing in the Era of Deep Learning: A Survey

Natural language is inherently a discrete symbolic representation of human knowledge. Recent advances in machine learning (ML) and in natural language processing (NLP) seem to contradict the above intuition: discrete symbols are fading away, erased by vectors or tensors called distributed and distributional representations. However, there is a strict link between distributed/distributional representations and discrete symbols, being the first an approximation of the second. A clearer understanding of the strict link between distributed/distributional representations and symbols may certainly lead to radically new deep learning networks. In this paper we make a survey that aims to renew the link between symbolic representations and distributed/distributional representations. This is the right time to revitalize the area of interpreting how discrete symbols are represented inside neural networks.

Cognitive Structures of Space-Time

In physics, the analysis of the space representing states of physical systems often takes the form of a layer-cake of increasingly rich structure. In this paper, we propose an analogous hierarchy in the cognition of spacetime. Firstly, we explore the interplay between the objective physical properties of space-time and the subjective compositional modes of relational representations within the reasoner. Secondly, we discuss the compositional structure within and between layers. The existing evidence in the available literature is reviewed to end with some testable consequences of our proposal at the brain and behavioral level.

From motor planning to execution: a sensorimotor loop perspective

How is an evanescent wish to move translated into a concrete action? This simple question and puzzling miracle remains a focal point of motor systems neuroscience. Where does the difficulty lie? A great deal has been known about biomechanics for quite some time. More recently, there have been significant advances in our understanding of how the spinal system is organized into modules corresponding to spinal synergies, which are fixed patterns of multimuscle recruitment. But much less is known about how the supraspinal system recruits these synergies in the correct spatiotemporal pattern to effectively control movement. We argue that what makes the problem of supraspinal control so difficult is that it emerges as a result of multiple convergent and redundant sensorimotor loops. Because these loops are convergent, multiple modes of information are mixed before being sent to the spinal system; because they are redundant, information is overlapping such that a mechanism must exist to eliminate the redundancy before the signal is sent to the spinal system. Given these complex interactions, simple correlation analyses between movement variables and neural activity are likely to render a confusing and inconsistent picture. Here, we suggest that the perspective of sensorimotor loops might help in achieving a better systems-level understanding. Furthermore, state-of-the-art techniques in neurotechnology, such as optogenetics, appear to be well suited for investigating the problem of motor control at the level of loops.

Two Ways to Build a Thought: Distinct Forms of Compositional Semantic Representation across Brain Regions

To understand a simple sentence such as "the woman chased the dog", the human mind must dynamically organize the relevant concepts to represent who did what to whom. This structured recombination of concepts (woman, dog, chased) enables the representation of novel events, and is thus a central feature of intelligence. Here, we use functional magnetic resonance (fMRI) and encoding models to delineate the contributions of three brain regions to the representation of relational combinations. We identify a region of anterior-medial prefrontal cortex (amPFC) that shares representations of noun-verb conjunctions across sentences: for example, a combination of "woman" and "chased" to encode woman-as-chaser, distinct from woman-as-chasee. This PFC region differs from the left-mid superior temporal cortex (lmSTC) and hippocampus, two regions previously implicated in representing relations. lmSTC represents broad role combinations that are shared across verbs (e.g., woman-as-agent), rather than narrow roles, limited to specific actions (woman-as-chaser). By contrast, a hippocampal sub-region represents events sharing narrow conjunctions as dissimilar. The success of the hippocampal conjunctive encoding model is anti-correlated with generalization performance in amPFC on a trial-by-trial basis, consistent with a pattern separation mechanism. Thus, these three regions appear to play distinct, but complementary, roles in encoding compositional event structure.

Linguistic generalization and compositionality in modern artificial neural networks

In the last decade, deep artificial neural networks have achieved astounding performance in many natural language-processing tasks. Given the high productivity of language, these models must possess effective generalization abilities. It is widely assumed that humans handle linguistic productivity by means of algebraic compositional rules: are deep networks similarly compositional? After reviewing the main innovations characterizing current deep language-processing networks, I discuss a set of studies suggesting that deep networks are capable of subtle grammar-dependent generalizations, but also that they do not rely on systematic compositional rules. I argue that the intriguing behaviour of these devices (still awaiting a full understanding) should be of interest to linguists and cognitive scientists, as it offers a new perspective on possible computational strategies to deal with linguistic productivity beyond rule-based compositionality, and it might lead to new insights into the less systematic generalization patterns that also appear in natural language. This article is part of the theme issue 'Towards mechanistic models of meaning composition'.

Tensors and compositionality in neural systems

Neither neurobiological nor process models of meaning composition specify the operator through which constituent parts are bound together into compositional structures. In this paper, we argue that a neurophysiological computation system cannot achieve the compositionality exhibited in human thought and language if it were to rely on a multiplicative operator to perform binding, as the tensor product (TP)-based systems that have been widely adopted in cognitive science, neuroscience and artificial intelligence do. We show via simulation and two behavioural experiments that TPs violate variable-value independence, but human behaviour does not. Specifically, TPs fail to capture that in the statements fuzzy cactus and fuzzy penguin, both cactus and penguin are predicated by fuzzy(x) and belong to the set of fuzzy things, rendering these arguments similar to each other. Consistent with that thesis, people judged arguments that shared the same role to be similar, even when those arguments themselves (e.g., cacti and penguins) were judged to be dissimilar when in isolation. By contrast, the similarity of the TPs representing fuzzy(cactus) and fuzzy(penguin) was determined by the similarity of the arguments, which in this case approaches zero. Based on these results, we argue that neural systems that use TPs for binding cannot approximate how the human mind and brain represent compositional information during processing. We describe a contrasting binding mechanism that any physiological or artificial neural system could use to maintain independence between a role and its argument, a prerequisite for compositionality and, thus, for instantiating the expressive power of human thought and language in a neural system. This article is part of the theme issue 'Towards mechanistic models of meaning composition'.

Modelling meaning composition from formalism to mechanism

Human thought and language have extraordinary expressive power because meaningful parts can be assembled into more complex semantic structures. This partly underlies our ability to compose meanings into endlessly novel configurations, and sets us apart from other species and current computing devices. Crucially, human behaviour, including language use and linguistic data, indicates that composing parts into complex structures does not threaten the existence of constituent parts as independent units in the system: parts and wholes exist simultaneously yet independently from one another in the mind and brain. This independence is evident in human behaviour, but it seems at odds with what is known about the brain's exquisite sensitivity to statistical patterns: everyday language use is productive and expressive precisely because it can go beyond statistical regularities. Formal theories in philosophy and linguistics explain this fact by assuming that language and thought are compositional: systems of representations that separate a variable (or role) from its values (fillers), such that the meaning of a complex expression is a function of the values assigned to the variables. The debate on whether and how compositional systems could be implemented in minds, brains and machines remains vigorous. However, it has not yet resulted in mechanistic models of semantic composition: how, then, are the constituents of thoughts and sentences put and held together? We review and discuss current efforts at understanding this problem, and we chart possible routes for future research. This article is part of the theme issue 'Towards mechanistic models of meaning composition'.

Trying to make it work: Compositional effects in the processing of compound "nonwords"

Speakers of languages with synchronically productive compounding systems, such as English, are likely to encounter new compounds on a daily basis. These can only be useful for communication if speakers are able to rapidly compose their meanings. However, while compositional meanings can be obtained for some novel compounds such as bridgemill, this is far harder for others such as radiosauce; accordingly, processing speed should be affected by the ease of such a compositional process. To rigorously test this hypothesis, we employed a fully implemented computational model based on distributional semantics to quantitatively measure the degree of semantic compositionality of novel compounds. In two large-scale studies, we collected timed sensibility judgements and lexical decisions for hundreds of morphologically structured nonwords in English. Response times were predicted by the constituents' semantic contribution to the compositional process, with slower rejections for more compositional nonwords. We found no indication of a difference in these compositional effects between the tasks, suggesting that speakers automatically engage in a compositional process whenever they encounter morphologically structured stimuli, even when it is not required by the task at hand. Such compositional effects in the processing of novel compounds have important implications for studies that employ such stimuli as filler material or "nonwords," as response times for these items can differ greatly depending on their compositionality.

Composition is the Core Driver of the Language-selective Network

The frontotemporal language network responds robustly and selectively to sentences. But the features of linguistic input that drive this response and the computations that these language areas support remain debated. Two key features of sentences are typically confounded in natural linguistic input: words in sentences (a) are semantically and syntactically combinable into phrase- and clause-level meanings, and (b) occur in an order licensed by the language's grammar. Inspired by recent psycholinguistic work establishing that language processing is robust to word order violations, we hypothesized that the core linguistic computation is composition, and, thus, can take place even when the word order violates the grammatical constraints of the language. This hypothesis predicts that a linguistic string should elicit a sentence-level response in the language network provided that the words in that string can enter into dependency relationships as in typical sentences. We tested this prediction across two fMRI experiments (total N = 47) by introducing a varying number of local word swaps into naturalistic sentences, leading to progressively less syntactically well-formed strings. Critically, local dependency relationships were preserved because combinable words remained close to each other. As predicted, word order degradation did not decrease the magnitude of the blood oxygen level-dependent response in the language network, except when combinable words were so far apart that composition among nearby words was highly unlikely. This finding demonstrates that composition is robust to word order violations, and that the language regions respond as strongly as they do to naturalistic linguistic input, providing that composition can take place.

The acquisition of compositional meaning

How do people create meaning from a string of sounds or pattern of dots? Insights into this process can be obtained from the way children acquire sentence meanings. According to the well-known principle of compositionality, the meaning of an expression is a function of the meanings of its parts and the way they are syntactically combined. However, children frequently seem to ignore syntactic structure in their sentence interpretations, suggesting that syntax is merely one of the sources of information constraining meaning and does not have a special status. A fundamental assumption in the argument in favour of compositionality is that speakers and listeners generally agree upon the meanings of sentences. Remarkably, however, children as listeners do not always understand what they are able to produce as speakers, and vice versa. For example, children's production of word order appears to develop ahead of their comprehension of word order in the acquisition of languages like English and Dutch. Such production-comprehension asymmetries are not uncommon in child language and motivate a view of compositionality as a principle pertaining to the result of perspective taking, and of meaning composition as a process of speaker-listener coordination. This article is part of the theme issue 'Towards mechanistic models of meaning composition'.

Phase synchronization varies systematically with linguistic structure composition

Computation in neuronal assemblies is putatively reflected in the excitatory and inhibitory cycles of activation distributed throughout the brain. In speech and language processing, coordination of these cycles resulting in phase synchronization has been argued to reflect the integration of information on different timescales (e.g. segmenting acoustics signals to phonemic and syllabic representations; (Giraud and Poeppel 2012 Nat. Neurosci. 15, 511 (doi:10.1038/nn.3063)). A natural extension of this claim is that phase synchronization functions similarly to support the inference of more abstract higher-level linguistic structures (Martin 2016 Front. Psychol. 7, 120; Martin and Doumas 2017 PLoS Biol. 15, e2000663 (doi:10.1371/journal.pbio.2000663); Martin and Doumas. 2019 Curr. Opin. Behav. Sci. 29, 77-83 (doi:10.1016/j.cobeha.2019.04.008)). Hale et al. (Hale et al. 2018 Finding syntax in human encephalography with beam search. arXiv 1806.04127 (http://arxiv.org/abs/1806.04127)) showed that syntactically driven parsing decisions predict electroencephalography (EEG) responses in the time domain; here we ask whether phase synchronization in the form of either inter-trial phrase coherence or cross-frequency coupling (CFC) between high-frequency (i.e. gamma) bursts and lower-frequency carrier signals (i.e. delta, theta), changes as the linguistic structures of compositional meaning (viz., bracket completions, as denoted by the onset of words that complete phrases) accrue. We use a naturalistic story-listening EEG dataset from Hale et al. to assess the relationship between linguistic structure and phase alignment. We observe increased phase synchronization as a function of phrase counts in the delta, theta, and gamma bands, especially for function words. A more complex pattern emerged for CFC as phrase count changed, possibly related to the lack of a one-to-one mapping between 'size' of linguistic structure and frequency band-an assumption that is tacit in recent frameworks. These results emphasize the important role that phase synchronization, desynchronization, and thus, inhibition, play in the construction of compositional meaning by distributed neural networks in the brain. This article is part of the theme issue 'Towards mechanistic models of meaning composition'.

Sheaving-a universal construction for semantic compositionality

Semantic compositionality—the way that meanings of complex entities obtain from meanings of constituent entities and their structural relations—is supposed to explain certain concomitant cognitive capacities, such as systematicity. Yet, cognitive scientists are divided on mechanisms for compositionality: e.g. a language of thought on one side versus a geometry of thought on the other. Category theory is a field of (meta)mathematics invented to bridge formal divides. We focus on sheaving—a construction at the nexus of algebra and geometry/topology, alluding to an integrative view, to sketch out a category theory perspective on the semantics of compositionality. Sheaving is a universal construction for making inferences from local knowledge, where meaning is grounded by the underlying topological space. Three examples illustrate how topology conveys meaning, in terms of the inclusion relations between the open sets that constitute the space, though the topology is not regarded as the only source of semantic information. In this sense, category (sheaf) theory provides a general framework for semantic compositionality.

This article is part of the theme issue ‘Towards mechanistic models of meaning composition’.

The syntax-semantics interface in animal vocal communication

Syntax (rules for combining words or elements) and semantics (meaning of expressions) are two pivotal features of human language, and interaction between them allows us to generate a limitless number of meaningful expressions. While both features were traditionally thought to be unique to human language, research over the past four decades has revealed intriguing parallels in animal communication systems. Many birds and mammals produce specific calls with distinct meanings, and some species combine multiple meaningful calls into syntactically ordered sequences. However, it remains largely unclear whether, like phrases or sentences in human language, the meaning of these call sequences depends on both the meanings of the component calls and their syntactic order. Here, leveraging recently demonstrated examples of meaningful call combinations, we introduce a framework for exploring the interaction between syntax and semantics (i.e. the syntax-semantic interface) in animal vocal sequences. We outline methods to test the cognitive mechanisms underlying the production and perception of animal vocal sequences and suggest potential evolutionary scenarios for syntactic communication. We hope that this review will stimulate phenomenological studies on animal vocal sequences as well as experimental studies on the cognitive processes, which promise to provide further insights into the evolution of language. This article is part of the theme issue 'What can animal communication teach us about human language?'

Microbiome Multi-Omics Network Analysis: Statistical Considerations, Limitations, and Opportunities

The advent of large-scale microbiome studies affords newfound analytical opportunities to understand how these communities of microbes operate and relate to their environment. However, the analytical methodology needed to model microbiome data and integrate them with other data constructs remains nascent. This emergent analytical toolset frequently ports over techniques developed in other multi-omics investigations, especially the growing array of statistical and computational techniques for integrating and representing data through networks. While network analysis has emerged as a powerful approach to modeling microbiome data, oftentimes by integrating these data with other types of omics data to discern their functional linkages, it is not always evident if the statistical details of the approach being applied are consistent with the assumptions of microbiome data or how they impact data interpretation. In this review, we overview some of the most important network methods for integrative analysis, with an emphasis on methods that have been applied or have great potential to be applied to the analysis of multi-omics integration of microbiome data. We compare advantages and disadvantages of various statistical tools, assess their applicability to microbiome data, and discuss their biological interpretability. We also highlight on-going statistical challenges and opportunities for integrative network analysis of microbiome data.

Toward accurate species-level metabarcoding of arthropod communities from the tropical forest canopy

Metabarcoding of arthropod communities can be used for assessing species diversity in tropical forests but the methodology requires validation for accurate and repeatable species occurrences in complex mixtures. This study investigates how the composition of ecological samples affects the accuracy of species recovery.Starting with field-collected bulk samples from the tropical canopy, the recovery of specimens was tested for subsets of different body sizes and major taxa, by assembling these subsets into increasingly complex composite pools. After metabarcoding, we track whether richness, diversity, and most importantly composition of any size class or taxonomic subset are affected by the presence of other subsets in the mixture.Operational taxonomic units (OTUs) greatly exceeded the number of morphospecies in most taxa, even under very stringent sequencing read filtering. There was no significant effect on the recovered OTU richness of small and medium-sized arthropods when metabarcoded alongside larger arthropods, despite substantial biomass differences in the mixture. The recovery of taxonomic subsets was not generally influenced by the presence of other taxa, although with some exceptions likely due to primer mismatches. Considerable compositional variation within size and taxon-based subcommunities was evident resulting in high beta-diversity among samples from within a single tree canopy, but this beta-diversity was not affected by experimental manipulation.We conclude that OTU recovery in complex arthropod communities, with sufficient sequencing depth and within reasonable size ranges, is not skewed by variable biomass of the constituent species. This could remove the need for time-intensive manual sorting prior to metabarcoding. However, there remains a chance of taxonomic bias, which may be primer-dependent. There will never be a panacea primer; instead, metabarcoding studies should carefully consider whether the aim is broadscale turnover, in which case these biases may not be important, or species lists, in which case separate PCRs and sequencing might be necessary. OTU number inflation remains an issue in metabarcoding and requires bioinformatic development, particularly in read filtering and OTU clustering, and/or greater use of species-identifying sequences generated outside of bulk sequencing.

Constructing Complexity in a Young Sign Language

A universally acknowledged, core property of language is its complexity, at each level of structure – sounds, words, phrases, clauses, utterances, and higher levels of discourse. How does this complexity originate and develop in a language? We cannot fully answer this question from spoken languages, since they are all thousands of years old or descended from old languages. However, sign languages of deaf communities can arise at any time and provide empirical data for testing hypotheses related to the emergence of language complexity. An added advantage of the signed modality is a correspondence between visible physical articulations and linguistic structures, providing a more transparent view of linguistic complexity and its emergence (Sandler, 2012). These essential characteristics of sign languages allow us to address the issue of emerging complexity by documenting the use of the body for linguistic purposes. We look at three types of discourse relations of increasing complexity motivated by research on spoken languages – additive, symmetric, and asymmetric (Mann and Thompson, 1988; Sanders et al., 1992). Each relation type can connect units at two different levels: within propositions (simpler) and across propositions (more complex).¹ We hypothesized that these relations provide a measure for charting the time course of emergence of complexity, from simplest to most complex, in a new sign language. We test this hypothesis on Israeli Sign Language (ISL), a young language, some of whose earliest users are still available for recording. Taking advantage of the unique relation in sign languages between bodily articulations and linguistic form, we study fifteen ISL signers from three generations, and demonstrate that the predictions indeed hold. We also find that younger signers tend to converge on more systematic marking of relations, that they use fewer articulators for a given linguistic function than older signers, and that the form of articulations becomes reduced, as the language matures. Mapping discourse relations to the bodily expression of linguistic components across age groups reveals how simpler, less constrained, and more gesture-like expressions, become language.

The Body as Evidence for the Nature of Language

Taking its cue from sign languages, this paper proposes that the recruitment and composition of body actions provide evidence for key properties of language and its emergence. Adopting the view that compositionality is the fundamental organizing property of language, we show first that actions of the hands, face, head, and torso in sign languages directly reflect linguistic components, and illuminate certain aspects of compositional organization among them that are relevant for all languages, signed and spoken. Studies of emerging sign languages strengthen the approach by showing that the gradual recruitment of bodily articulators for linguistic functions directly maps the way in which a new language increases in complexity and efficiency over time. While compositional communication is almost exclusively restricted to humans, it is not restricted to language. In the spontaneous, intense emotional displays of athletes, different emotional states are correlated with actions of particular face and body features and feature groupings. These findings indicate a much more ancient communicative compositional capacity, and support a paradigm that includes visible body actions in the quest for core linguistic properties and their origins.

Wild Birds Use an Ordering Rule to Decode Novel Call Sequences

The generative power of human language depends on grammatical rules, such as word ordering, that allow us to produce and comprehend even novel combinations of words [1-3]. Several species of birds and mammals produce sequences of calls [4-6], and, like words in human sentences, their order may influence receiver responses [7]. However, it is unknown whether animals use call ordering to extract meaning from truly novel sequences. Here, we use a novel experimental approach to test this in a wild bird species, the Japanese tit (Parus minor). Japanese tits are attracted to mobbing a predator when they hear conspecific alert and recruitment calls ordered as alert-recruitment sequences [7]. They also approach in response to recruitment calls of heterospecific individuals in mixed-species flocks [8, 9]. Using experimental playbacks, we assess their responses to artificial sequences in which their own alert calls are combined into different orderings with heterospecific recruitment calls. We find that Japanese tits respond similarly to mixed-species alert-recruitment call sequences and to their own alert-recruitment sequences. Importantly, however, tits rarely respond to mixed-species sequences in which the call order is reversed. Thus, Japanese tits extract a compound meaning from novel call sequences using an ordering rule. These results demonstrate a new parallel between animal communication systems and human language, opening new avenues for exploring the evolution of ordering rules and compositionality in animal vocal sequences.

Balance Trees Reveal Microbial Niche Differentiation

Advances in sequencing technologies have enabled novel insights into microbial niche differentiation, from analyzing environmental samples to understanding human diseases and informing dietary studies. However, identifying the microbial taxa that differentiate these samples can be challenging. These issues stem from the compositional nature of 16S rRNA gene data (or, more generally, taxon or functional gene data); the changes in the relative abundance of one taxon influence the apparent abundances of the others. Here we acknowledge that inferring properties of individual bacteria is a difficult problem and instead introduce the concept of balances to infer meaningful properties of subcommunities, rather than properties of individual species. We show that balances can yield insights about niche differentiation across multiple microbial environments, including soil environments and lung sputum. These techniques have the potential to reshape how we carry out future ecological analyses aimed at revealing differences in relative taxonomic abundances across different samples. IMPORTANCE By explicitly accounting for the compositional nature of 16S rRNA gene data through the concept of balances, balance trees yield novel biological insights into niche differentiation. The software to perform this analysis is available under an open-source license and can be obtained at https://github.com/biocore/gneiss. Author Video: An author video summary of this article is available.

The Neuronal Correlates of Indeterminate Sentence Comprehension: An fMRI Study

Sentences such as The author started the book are indeterminate because they do not make explicit what the subject (the author) started doing with the object (the book). In principle, indeterminate sentences allow for an infinite number of interpretations. One theory, however, assumes that these sentences are resolved by semanticcoercion, a linguistic process that forces the noun book to be interpreted as an activity (e.g., writing the book) or by a process that interpolates this activity information in the resulting enriched semantic composition. An alternative theory, pragmatic, assumes classical semantic composition, whereby meaning arises from the denotation of words and how they are combined syntactically, with enrichment obtained via pragmatic inferences beyond linguistic-semantic processes. Cognitive neuroscience studies investigating the neuroanatomical and functional correlates of indeterminate sentences have shown activations either at the ventromedial pre-frontal cortex (vmPFC) or at the left inferior frontal gyrus (L-IFG). These studies have supported the semantic coercion theory assuming that one of these regions is where enriched semantic composition takes place. Employing functional magnetic resonance imaging (fMRI), we found that indeterminate sentences activate bilaterally the superior temporal gyrus (STG), the right inferior frontal gyrus (R-IFG), and the anterior cingulate cortex (ACC), more so than control sentences (The author wrote the book). Activation of indeterminate sentences exceeded that of anomalous sentences (…drank the book) and engaged more left- and right-hemisphere areas than other sentence types. We suggest that the widespread activations for indeterminate sentences represent the deployment of pragmatic-inferential processes, which seek to enrich sentence content without necessarily resorting to semantic coercion.

Neural correlate of the construction of sentence meaning

The neural processes that underlie your ability to read and understand this sentence are unknown. Sentence comprehension occurs very rapidly, and can only be understood at a mechanistic level by discovering the precise sequence of underlying computational and neural events. However, we have no continuous and online neural measure of sentence processing with high spatial and temporal resolution. Here we report just such a measure: intracranial recordings from the surface of the human brain show that neural activity, indexed by γ-power, increases monotonically over the course of a sentence as people read it. This steady increase in activity is absent when people read and remember nonword-lists, despite the higher cognitive demand entailed, ruling out accounts in terms of generic attention, working memory, and cognitive load. Response increases are lower for sentence structure without meaning ("Jabberwocky" sentences) and word meaning without sentence structure (word-lists), showing that this effect is not explained by responses to syntax or word meaning alone. Instead, the full effect is found only for sentences, implicating compositional processes of sentence understanding, a striking and unique feature of human language not shared with animal communication systems. This work opens up new avenues for investigating the sequence of neural events that underlie the construction of linguistic meaning.

Why Are There Failures of Systematicity? The Empirical Costs and Benefits of Inducing Universal Constructions

Systematicity is a property of cognition where capacity for certain cognitive abilities implies capacity for certain other (structurally related) cognitive abilities. This property is thought to derive from a capacity to represent/process common structural relations between constituents of cognizable entities, however, systematicity may not always materialize in such admissible contexts. A theoretical challenge is to explain why systematicity fails to materialize in contexts that allow the realization (e.g., by induction) of common structure (universal construction). We hypothesize that one cause of failure arises when the potential gain afforded by induction of common structure is overshadowed by the immediate benefit of learning the task as independent stimulus-response associations. This hypothesis was tested in an experiment that required learning two series of pair maps that involved products (universal construction), or non-products (control) of varied size: the number of unique cue/target elements (three to six) constituting pairs. Each series was learned in either ascending or descending order of size. Only performance on the product series was affected by order: systematicity was obtained universally in the descend group, but only on large sets in the ascend group, as revealed by the significant order × size interaction for errors in the product condition, F (3, 87) = 3.38, p < 0.05. Smaller maps are more easily learned without inducing the common product structure, which is more readily observable with larger maps: larger maps provide more evidence for relationships between stimulus dimensions that facilitate the discovery of the common structure. The new challenge, then, is to explain the systematic learnability of stimulus-response maps, i.e., second-order systematicity.

Causal Evidence for a Mechanism of Semantic Integration in the Angular Gyrus as Revealed by High-Definition Transcranial Direct Current Stimulation

A defining aspect of human cognition is the ability to integrate conceptual information into complex semantic combinations. For example, we can comprehend "plaid" and "jacket" as individual concepts, but we can also effortlessly combine these concepts to form the semantic representation of "plaid jacket." Many neuroanatomic models of semantic memory propose that heteromodal cortical hubs integrate distributed semantic features into coherent representations. However, little work has specifically examined these proposed integrative mechanisms and the causal role of these regions in semantic integration. Here, we test the hypothesis that the angular gyrus (AG) is critical for integrating semantic information by applying high-definition transcranial direct current stimulation (tDCS) to an fMRI-guided region-of-interest in the left AG. We found that anodal stimulation to the left AG modulated semantic integration but had no effect on a letter-string control task. Specifically, anodal stimulation to the left AG resulted in faster comprehension of semantically meaningful combinations like "tiny radish" relative to non-meaningful combinations, such as "fast blueberry," when compared to the effects observed during sham stimulation and stimulation to a right-hemisphere control brain region. Moreover, the size of the effect from brain stimulation correlated with the degree of semantic coherence between the word pairs. These findings demonstrate that the left AG plays a causal role in the integration of lexical-semantic information, and that high-definition tDCS to an associative cortical hub can selectively modulate integrative processes in semantic memory.

SIGNIFICANCE STATEMENT

A major goal of neuroscience is to understand the neural basis of behaviors that are fundamental to human intelligence. One essential behavior is the ability to integrate conceptual knowledge from semantic memory, allowing us to construct an almost unlimited number of complex concepts from a limited set of basic constituents (e.g., "leaf" and "wet" can be combined into the more complex representation "wet leaf"). Here, we present a novel approach to studying integrative processes in semantic memory by applying focal brain stimulation to a heteromodal cortical hub implicated in semantic processing. Our findings demonstrate a causal role of the left angular gyrus in lexical-semantic integration and provide motivation for novel therapeutic applications in patients with lexical-semantic deficits.

Systematicity and a Categorical Theory of Cognitive Architecture: Universal Construction in Context

Why does the capacity to think certain thoughts imply the capacity to think certain other, structurally related, thoughts? Despite decades of intensive debate, cognitive scientists have yet to reach a consensus on an explanation for this property of cognitive architecture—the basic processes and modes of composition that together afford cognitive capacity—called systematicity. Systematicity is generally considered to involve a capacity to represent/process common structural relations among the equivalently cognizable entities. However, the predominant theoretical approaches to the systematicity problem, i.e., classical (symbolic) and connectionist (subsymbolic), require arbitrary (ad hoc) assumptions to derive systematicity. That is, their core principles and assumptions do not provide the necessary and sufficient conditions from which systematicity follows, as required of a causal theory. Hence, these approaches fail to fully explain why systematicity is a (near) universal property of human cognition, albeit in restricted contexts. We review an alternative, category theory approach to the systematicity problem. As a mathematical theory of structure, category theory provides necessary and sufficient conditions for systematicity in the form of universal construction: each systematically related cognitive capacity is composed of a common component and a unique component. Moreover, every universal construction can be viewed as the optimal construction in the given context (category). From this view, universal constructions are derived from learning, as an optimization. The ultimate challenge, then, is to explain the determination of context. If context is a category, then a natural extension toward addressing this question is higher-order category theory, where categories themselves are the objects of construction.