Peter Zeng, Weiling Li, Amie J. Paige, Zhengxiang Wang, Panagiotis Kaliosis, Dimitris Samaras, Gregory J. Zelinsky, Susan Brennan, Owen Rambow. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

In this paper, we examine whether LLMs can generate context-appropriate word order in Russian. Using a new corpus of movie scripts in Russian, with a particular focus on dialogues, we investigate how various semantic, morphosyntactic, and discourse features influence word order choice. We show that traditional machine learning can use these features to model different word order fairly well. We also examine whether LLM dialogue partners can generate context-appropriate word order in Russian. With both zero- and few-shot prompting, LLMs fail to generate word orders beyond the default subject-verb-object. Instead, in order to generate context-appropriate word orders, LLMs need to be fine-tuned or given explicit word order suggestions from a traditional machine learning method.

Deep learning models excel in various Natural Language Processing (NLP) tasks, but their performance (excluding approaches like zero-shot learning or few-shot learning) relies on ample data, posing challenges in fields with limited datasets. To address the poverty in the size of training data, a number of approaches could be taken, such as multi-task learning and data augmentation. Aiming to leverage Large Language Models (LLMs), we propose a data augmentation algorithm. It subtly alters sentences by inserting random words and utilizes LLMs to find the most fitting replacements within their embedding space. Taking inspiration from Prompt Tuning, the focus shifts from optimizing the input prompt to updating the inserted tokens’ embedding vectors by maximizing the conditional generation probability. This allows for vast sample generation while implicitly benefiting from the knowledge within LLMs. The results from our extensive set of experiments on various benchmark text classification tasks show a substantial improvement over the non-augmented outcomes.

The paper explores the performance of LLMs in the context of multi-dimensional analytic writing assessments, i.e. their ability to provide both scores and comments based on multiple assessment criteria.Using a corpus of literature reviews written by L2 graduate students and assessed by human experts against 9 analytic criteria, we prompt several popular LLMs to perform the same task under various conditions.To evaluate the quality of feedback comments, we apply a novel feedback comment quality evaluation framework.This framework is interpretable, cost-efficient, scalable, and reproducible, compared to existing methods that rely on manual judgments.We find that LLMs can generate reasonably good and generally reliable multi-dimensional analytic assessments.We release our corpus and code 1 for reproducibility.

During conversation, speakers collaborate on spontaneous referring expressions, which they can then re-use in subsequent conversation with the same partner.Understanding such referring expressions is an important ability for an embodied agent so that it can carry out tasks in the real world.This requires integrating and understanding language, vision, and conversational interaction.We study the capabilities of seven state-of-the-art Large Vision Language Models (LVLMs) as overhearers to a corpus of spontaneous conversations between pairs of human discourse participants engaged in a collaborative object-matching task.We find that such a task remains challenging for current LVLMs, which fail to show a consistent performance improvement as they overhear more conversations from the same discourse participants repeating the same task for multiple rounds.We release our corpus and code 1 for reproducibility and to facilitate future research.

Saeed Ahmadnia, Arash Yousefi Jordehi, Mahsa Hosseini Khasheh Heyran, Seyed Abolghasem Mirroshandel, Owen Rambow, Cornelia Caragea. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Opinion mining is an important task in natural language processing. The MPQA Opinion Corpus is a fine-grained and comprehensive dataset of private states (i.e., the condition of a source who has an attitude which may be directed toward a target) based on context. Although this dataset was released years ago, because of its complex definition of annotations and hard-to-read data format, almost all existing research works have only focused on a small subset of the dataset. In this paper, we present a comprehensive study of the entire MPQA 2.0 dataset. In order to achieve this goal, we first provide a clean version of MPQA 2.0 in a more interpretable format. Then, we propose two novel approaches for opinion mining, establishing new high baselines for future work. We use two pre-trained large language models, BERT and T5, to automatically identify the type, polarity, and intensity of private states expressed in phrases, and we use T5 to detect opinion expressions and their agents (i.e., sources).

The notion of face described by Brown and Levinson (1987) has been studied in great detail, but a critical aspect of the framework, that which focuses on how intentions mediate the planning of turns which impose upon face, has received far less attention. We present an analysis of three computational systems trained for classifying both intention and politeness, focusing on how the former influences the latter. In politeness theory, agents attend to the desire to have their wants appreciated (positive face), and a complementary desire to act unimpeded and maintain freedom (negative face). Similar to speech acts, utterances can perform so-called face acts which can either raise or threaten the positive or negative face of the speaker or hearer. We begin by using an existing corpus to train a model which classifies face acts, achieving a new SoTA in the process. We then observe that every face act has an underlying intention that motivates it and perform additional experiments integrating dialog act annotations to provide these intentions by proxy. Our analysis finds that dialog acts improve performance on face act detection for minority classes and points to a close relationship between aspects of face and intent.

Hedges allow speakers to mark utterances as provisional, whether to signal nonprototypicality or "fuzziness", to indicate a lack of commitment to an utterance, to attribute responsibility for a statement to someone else, to invite input from a partner, or to soften critical feedback in the service of face-management needs.Here we focus on hedges in an experimentally parameterized corpus of 63 Roadrunner cartoon narratives spontaneously produced from memory by 21 speakers for copresent addressees, transcribed to text (Galati and Brennan, 2010).We created a gold standard of hedges annotated by human coders (the Roadrunner-Hedge corpus) and compared three LLM-based approaches for hedge detection: fine-tuning BERT, and zero and fewshot prompting with GPT-4o and LLaMA-3.The best-performing approach was a fine-tuned BERT model, followed by few-shot GPT-4o.After an error analysis on the top performing approaches, we used an LLM-in-the-Loop approach to improve the gold standard coding, as well as to highlight cases in which hedges are ambiguous in linguistically interesting ways that will guide future research.This is the first step in our research program to train LLMs to interpret and generate collateral signals appropriately and meaningfully in conversation.

We propose a framework for analyzing discourse by combining two interdependent concepts from sociolinguistic theory: face acts and politeness.While politeness has robust existing tools and data, face acts are less resourced.We introduce a new corpus created by annotating Wikipedia talk pages with face acts and we use this to train a face act tagger.We then employ our framework to study how face and politeness interact with gender and power in discussions between Wikipedia editors.Among other findings, we observe that female Wikipedians are not only more polite, which is consistent with prior studies, but that this difference corresponds with significantly more language directed at humbling aspects of their own face.Interestingly, the distinction nearly vanishes once limiting to editors with administrative power.

Explicit linguistic knowledge, encoded by resources such as rule-based morphological analyzers, continues to prove useful in downstream NLP tasks, especially for low-resource languages and dialects.Rules are an important asset in descriptive linguistic grammars.However, creating such resources is usually expensive and non-trivial, especially for spoken varieties with no written standard.In this work, we present a novel approach for automatically learning morphophonological rules of Arabic from a corpus.Motivated by classic cognitive models for rule learning, rules are generalized cautiously.Rules that are memorized for individual items are only allowed to generalize to unseen forms if they are sufficiently reliable in the training data.The learned rules are further examined to ensure that they capture true linguistic phenomena described by domain experts.We also investigate the learnability of rules in low-resource settings across different experimental setups and dialects

Knowledge of norms is needed to understand and reason about acceptable behavior in human communication and interactions across sociocultural scenarios.Most computational research on norms has focused on a single culture, and manually built datasets, from nonconversational settings.We address these limitations by proposing a new framework, NORMSAGE 1 , to automatically extract culturespecific norms from multi-lingual conversations.NORMSAGE uses GPT-3 prompting to 1) extract candidate norms directly from conversations and 2) provide explainable selfverification to ensure correctness and relevance.Comprehensive empirical results show the promise of our approach to extract highquality culture-aware norms from multi-lingual conversations (English and Chinese), across several quality metrics.Further, our relevance verification can be extended to assess the adherence and violation of any norm with respect to a conversation on-the-fly, along with textual explanation.NORMSAGE achieves an AUC of 94.6% in this grounding setup, with generated explanations matching human-written quality.𝐚) 𝐈𝐧𝐢𝐭𝐢𝐚𝐥 𝐃𝐢𝐬𝐜𝐨𝐯𝐞𝐫𝒚: 𝒅𝒗𝒓(⋅) irrelevant contradict entail Correctness Verdict ( ! 𝑪 𝒗 ): Correctness Explanation ( ! 𝑪 𝒆 ): Yes, honesty is the foundation of trust, and strong family relationships are built on trust.

We present a corrected version of a subset of the FactBank data set. Previously published results on FactBank are no longer valid. We perform experiments on FactBank using multiple training paradigms, data smoothing techniques, and polarity classifiers. We argue that f-measure is an important alternative evaluation metric for factuality. We provide new state-of-the-art results for four corpora including FactBank. We perform an error analysis on Factbank combined with two similar corpora.

We present the BeSt corpus, which records cognitive state: who believes what (i.e., factuality), and who has what sentiment towards what. This corpus is inspired by similar source-and-target corpora, specifically MPQA and FactBank. The corpus comprises two genres, newswire and discussion forums, in three languages, Chinese (Mandarin), English, and Spanish. The corpus is distributed through the LDC.

Many tasks aim to measure MACHINE READ-ING COMPREHENSION (MRC), often focusing on question types presumed to be difficult.Rarely, however, do task designers start by considering what systems should in fact comprehend.In this paper we make two key contributions.First, we argue that existing approaches do not adequately define comprehension; they are too unsystematic about what content is tested.Second, we present a detailed definition of comprehension-a TEM-PLATE OF UNDERSTANDING-for a widely useful class of texts, namely short narratives.We then conduct an experiment that strongly suggests existing systems are not up to the task of narrative understanding as we define it.

Existing methods for different document classification tasks in the context of social networks typically only capture the semantics of texts, while ignoring the users who exchange the text and the network they form. However, some work has shown that incorporating the social network information in addition to information from language is effective for various NLP applications including sentiment analysis, inferring user attributes, and predicting inter-personal relations. In this paper, we present an empirical study of email classification into “Business” and “Personal” categories. We represent the email communication using various graph structures. As features, we use both the textual information from the email content and social network information from the communication graphs. We also model the thread structure for emails. We focus on detecting personal emails, and we evaluate our methods on two corpora, only one of which we train on. The experimental results reveal that incorporating social network information improves over the performance of an approach based on textual information only. The results also show that considering the thread structure of emails improves the performance further. Furthermore, our approach improves over a state-of-the-art baseline which uses node embeddings based on both lexical and social network information.

Jungo Kasai, Robert Frank, Pauli Xu, William Merrill, Owen Rambow. Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers). 2018.

Vinodkumar Prabhakaran, Premkumar Ganeshkumar, Owen Rambow. Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers). 2018.

We present supertagging-based models for Tree Adjoining Grammar parsing that use neural network architectures and dense vector representation of supertags (elementary trees) to achieve state-of-the-art performance in unlabeled and labeled attachment scores.The shift-reduce parsing model eschews lexical information entirely, and uses only the 1-best supertags to parse a sentence, providing further support for the claim that supertagging is "almost parsing."We demonstrate that the embedding vector representations the parser induces for supertags possess linguistically interpretable structure, supporting analogies between grammatical structures like those familiar from recent work in distributional semantics.This dense representation of supertags overcomes the drawbacks for statistical models of TAG as compared to CCG parsing, raising the possibility that TAG is a viable alternative for NLP tasks that require the assignment of richer structural descriptions to sentences.

Violence is a serious problems for cities like Chicago and has been exacerbated by the use of social media by gang-involved youths for taunting rival gangs. We present a corpus of tweets from a young and powerful female gang member and her communicators, which we have annotated with discourse intention, using a deep read to understand how and what triggered conversations to escalate into aggression. We use this corpus to develop a part-of-speech tagger and phrase table for the variant of English that is used and a classifier for identifying tweets that express grieving and aggression.

Arabic dialects present a special problem for natural language processing because there are few resources, they have no standard orthography, and have not been studied much. However, as more and more written dialectal Arabic is found in social media, NLP for Arabic dialects becomes an important goal. We present a methodology for creating a morphological analyzer and a morphological tagger for dialectal Arabic, and we illustrate it on Egyptian and Levantine Arabic. To our knowledge, these are the first analyzer and tagger for Levantine.

We investigate using Adaptor Grammars for unsupervised morphological segmentation. Using six development languages, we investigate in detail different grammars, the use of morphological knowledge from outside sources, and the use of a cascaded architecture. Using cross-validation on our development languages, we propose a system which is language-independent. We show that it outperforms two state-of-the-art systems on 5 out of 6 languages.

We present experiments in incrementally learning a dependency parser. The parser will be used in the WordsEye Linguistics Tools (WELT) (Ulinski et al., 2014) which supports field linguists documenting a language’s syntax and semantics. Our goal is to make syntactic annotation faster for field linguists. We have created a new parallel corpus of descriptions of spatial relations and motion events, based on pictures and video clips used by field linguists for elicitation of language from native speaker informants. We collected descriptions for each picture and video from native speakers in English, Spanish, German, and Egyptian Arabic. We compare the performance of MSTParser (McDonald et al., 2006) and MaltParser (Nivre et al., 2006) when trained on small amounts of this data. We find that MaltParser achieves the best performance. We also present the results of experiments using the parser to assist with annotation. We find that even when the parser is trained on a single sentence from the corpus, annotation time significantly decreases.

Sentiment analysis has been a major area of interest, for which the existence of highquality resources is crucial.In Arabic, there is a reasonable number of sentiment lexicons but with major deficiencies.The paper presents a large-scale Standard Arabic Sentiment Lexicon (SLSA) that is publicly available for free and avoids the deficiencies in the current resources.SLSA has the highest up-to-date reported coverage.The construction of SLSA is based on linking the lexicon of AraMorph with Sen-tiWordNet along with a few heuristics and powerful back-off.SLSA shows a relative improvement of 37.8% over a state-of-theart lexicon when tested for accuracy.It also outperforms it by an absolute 3.5% of F1-score when tested for sentiment analysis.

We present a novel way of generating unseen words, which is useful for certain applications such as automatic speech recognition or optical character recognition in low-resource languages. We test our vocabulary generator on seven low-resource languages by measuring the decrease in out-of-vocabulary word rate on a held-out test set. The languages we study have very different morphological properties; we show how our results differ depending on the morphological complexity of the language. In our best result (on Assamese), our approach can predict 29% of the token-based out-of-vocabulary with a small amount of unlabeled training data.

In this paper, we address the problem of converting Dialectal Arabic (DA) text that is written in the Latin script (called Arabizi) into Arabic script following the CODA convention for DA orthography. The presented system uses a finite state transducer trained at the character level to generate all possible transliterations for the input Arabizi words. We then filter the generated list using a DA morpholog-ical analyzer. After that we pick the best choice for each input word using a lan-guage model. We achieve an accuracy of 69.4 % on an unseen test set compared to 63.1 % using a system which represents a previously proposed approach. 1

Apoorv Agarwal, Sriramkumar Balasubramanian, Anup Kotalwar, Jiehan Zheng, Owen Rambow. Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics. 2014.

We study the topic dynamics of interac-tions in political debates using the 2012 Republican presidential primary debates as data. We show that the tendency of candidates to shift topics changes over the course of the election campaign, and that it is correlated with their relative power. We also show that our topic shift features help predict candidates ’ relative rankings. 1

We investigate the interaction of power, gender, and language use in the Enron email corpus.We present a freely available extension to the Enron corpus, with the gender of senders of 87% messages reliably identified.Using this data, we test two specific hypotheses drawn from the sociolinguistic literature pertaining to gender and power: women managers use face-saving communicative strategies, and women use language more explicitly than men to create and maintain social relations.We introduce the notion of "gender environment" to the computational study of written conversations; we interpret this notion as the gender makeup of an email thread, and show that some manifestations of power differ significantly between gender environments.Finally, we show the utility of gender information in the problem of automatically predicting the direction of power between pairs of participants in email interactions.

We present a method for automatically learning inflectional classes and associated lemmas from morphologically annotated corpora.The method consists of a core languageindependent algorithm, which can be optimized for specific languages.The method is demonstrated on Egyptian Arabic and German, two morphologically rich languages.Our best method for Egyptian Arabic provides an error reduction of 55.6% over a simple baseline; our best method for German achieves a 66.7% error reduction.

We explore the contribution of lexical and inflectional morphology features to dependency parsing of Arabic, a morphologically rich language with complex agreement patterns. Using controlled experiments, we contrast the contribution of different part-of-speech (POS) tag sets and morphological features in two input conditions: machine-predicted condition (in which POS tags and morphological feature values are automatically assigned), and gold condition (in which their true values are known). We find that more informative (fine-grained) tag sets are useful in the gold condition, but may be detrimental in the predicted condition, where they are outperformed by simpler but more accurately predicted tag sets. We identify a set of features (definiteness, person, number, gender, and undiacritized lemma) that improve parsing quality in the predicted condition, whereas other features are more useful in gold. We are the first to show that functional features for gender and number (e.g., “broken plurals”), and optionally the related rationality (“humanness”) feature, are more helpful for parsing than form-based gender and number. We finally show that parsing quality in the predicted condition can dramatically improve by training in a combined gold+predicted condition. We experimented with two transition-based parsers, MaltParser and Easy-First Parser. Our findings are robust across parsers, models, and input conditions. This suggests that the contribution of the linguistic knowledge in the tag sets and features we identified goes beyond particular experimental settings, and may be informative for other parsers and morphologically rich languages.

Abstract This paper focuses on an important step in the creation of a system of meaning representation and the development of semantically annotated parallel corpora, for use in applications such as machine translation, question answering, text summarization, and information retrieval. The work described below constitutes the first effort of any kind to annotate multiple translations of foreign-language texts with interlingual content. Three levels of representation are introduced: deep syntactic dependencies (IL0), intermediate semantic representations (IL1), and a normalized representation that unifies conversives, nonliteral language, and paraphrase (IL2). The resulting annotated, multilingually induced, parallel corpora will be useful as an empirical basis for a wide range of research, including the development and evaluation of interlingual NLP systems and paraphrase-extraction systems as well as a host of other research and development efforts in theoretical and applied linguistics, foreign language pedagogy, translation studies, and other related disciplines.

We present MAGEAD, a morphological analyzer and generator for the Arabic language family.Our work is novel in that it explicitly addresses the need for processing the morphology of the dialects.MAGEAD performs an on-line analysis to or generation from a root+pattern+features representation, it has separate phonological and orthographic representations, and it allows for combining morphemes from different dialects.We present a detailed evaluation of MAGEAD.

We present an approach to using a morphological analyzer for tokenizing and morphologically tagging (including partof-speech tagging) Arabic words in one process.We learn classifiers for individual morphological features, as well as ways of using these classifiers to choose among entries from the output of the analyzer.We obtain accuracy rates on all tasks in the high nineties.

Many speech and language processing problems have been successfully cast as classification problems - associating a token with a label from a prespecified label set. However, in all these applications, the set of labels is regarded as a flat list of symbols with no inherent internal structure and no co-constraints among the labels. We present a classification task using structured word labels, called supertags, and discuss methods that exploit the structure of these labels in the context of natural language generation. Supertags encode predicate-argument information along with syntactic ordering constraints that can be exploited for realizing a sentence. We report the accuracy of supertagging models for a node using features derived from its local tree context and attributes of the supertags.

We present conditions under which verb phrases are elided based on a corpus of positive and negative examples. Factor that affect verb phrase ellipsis include: the distance between antecedent and ellipsis site, the syntactic relation between antecedent and ellipsis site, and the presence or absence of adjuncts. Building on these results, we examine where in the generation architecture a trainable algorithm for VP ellipsis should be located. We show that the best performance is achieved when the trainable module is located after the realizer and has access to surface-oriented features (error rate of 7.5%).

Techniques for automatically training modules of a natural language generator have recently been proposed, but a fundamental concern is whether the quality of utterances produced with trainable components can compete with hand-crafted template-based or rule-based approaches. In this paper We experimentally evaluate a trainable sentence planner for a spoken dialogue system by eliciting subjective human judgments. In order to perform an exhaustive comparison, we also evaluate a hand-crafted template-based generation component, two rule-based sentence planners, and two baseline sentence planners. We show that the trainable sentence planner performs better than the rule-based systems and the baselines, and as well as the hand-crafted system.

Despite the advantages that object technology can provide to the software development community and its customers, the fundamental problems associated with identifying objects, their attributes, and methods remain: it is a largely manual process driven by heuristics that analysts acquire through experience. While a number of methods exist for requirements development and specification, very few tools exist to assist analysts in making the transition from textual descriptions to other notations for object-oriented analysis and other conceptual models. We describe a methodology and a prototype tool. Linguistic Assistant for Domain Analysis (LIDA), which provide linguistic assistance in the model development process. We first present our methodology to conceptual modeling through linguistic analysis. We give an overview of LIDA's functionality and present its technical design and the functionality of its components. We also provide a comparison of LIDA's functionality with that of other research prototypes. Finally, we present an example of how LIDA is used in a conceptual modeling task.

We describe experiments designed to learn associations between two types of intonational features, pitch accent and phrasing, from a tree-based corpus annotated with various intonational and syntactic features, for a concept-to-speech system. We show that using novel tree-based features improves the quality of boundary prediction over using only the linear orderbased features normally used in text-to-speech.

The dialog manager of a spoken dialog system often performs domain dependent functions as well as general dialog tasks. It is possible to separate the domain specific knowledge from knowledge about language using techniques from natural language generation. However a natural language generator often has to be tuned for particular applications. In this work, we describe a new method for automatically training the natural language generator and examine the role that domain specific and domain independent features have on performance. We show that although the general features have the largest impact, the use of domain specific features improves performance, while still retaining the benefits of automatic domain customization through training. 1.

There is considerable interest among computational linguists in lexicalized grammatical frameworks; lexicalized tree adjoining grammar (LTAG) is one widely studied example. In this paper, we investigate how derivations in LTAG can be viewed not as manipulations of trees but as manipulations of tree descriptions. Changing the way the lexicalized formalism is viewed raises questions as to the desirability of certain aspects of the formalism. We present a new formalism, d-tree substitution grammar (DSG). Derivations in DSG involve the composition of d-trees, special kinds of tree descriptions. Trees are read off from derived d-trees. We show how the DSG formalism, which is designed to inherit many of the characterestics of LTAG, can be used to express a variety of linguistic analyses not available in LTAG.

Choosing the best lexeme to realize a meaning in natural language generation is a hard task. We investigate different tree-based stochastic models for lexical choice. Because of the difficulty of obtaining a sense-tagged corpus, we generalize the notion of synonymy. We show that a tree-based model can achieve a word-bag based accuracy of 90%, representing an improvement over the baseline.

Certain generation applications may profit from the use of stochastic methods. In developing stochastic methods, it is crucial to be able to quickly assess the relative merits of different approaches or models. In this paper, we present several types of intrinsic (system internal) metrics which we have used for baseline quantitative assessment. This quantitative assessment should then be augmented to a fuller evaluation that examines qualitative aspects. To this end, we describe an experiment that tests correlation between the quantitative metrics and human qualitative judgment. The experiment confirms that intrinsic metrics cannot replace human evaluation, but some correlate significantly with human judgments of quality and understandability and can be used for evaluation during development.

In synchronous rewriting, the productions of two rewriting systems are paired and applied synchronously in the derivation of a pair of strings. We present a new synchronous rewriting system and argue that it can handle certain phenomena that are not covered by existing synchronous systems. We also prove some interesting formal/computational properties of our system.

DTG are designed to share some of the advantages of TAG while overcoming some of its limitations. DTG involve two composition operations called subsertion and sister-adjunction. The most distinctive feature of DTG is that, unlike TAG, there is complete uniformity in the way that the two DTG operations relate lexical items: subsertion always corresponds to complementation and sister-adjunction to modification. Furthermore, DTG, unlike TAG, can provide a uniform analysis for wh-movement in English and Kashmiri, despite the fact that the wh element in Kashmiri appears in sentence-second position, and not sentence-initial position as in English.

This paper defines multiset-valued linear index grammar and unordered vector grammar with dominance links. The former models certain uses of multiset-valued feature structures in unification-based formalisms, while the latter is motivated by word order variation and by "quasi-trees", a generalization of trees. The two formalisms are weakly equivalent, and an important subset is at most context-sensitive and polynomially parsable.

In many head‐final languages such as German, Hindi, Japanese, and Korean, but also in some other languages such as Russian, arguments of a verb can occur in any order. Furthermore, arguments can occur outside of their clause (“long‐distance scrambling”). Long‐distance scrambling presents a challenge both to linguistic theory and to formal frameworks for linguistic description because it is very unconstrained: in a given sentence, there is no bound on the number of elements that can be scrambled nor on the distance over which each element can scramble. We discuss two formal frameworks related to tree‐adjoining grammar. First, we show how scrambling facts from Korean can be handled by nonlocal multicomponent TAG (MC‐TAG). Then, we argue that overt vWt‐movement in German makes this analysis unattractive, and suggest a new version of MC‐TAG, called V‐TAG, which can handle both Korean and German word order variation. Interestingly, this new version has more attractive computational properties than nonlocal MC‐TAG. We conclude that this formalism is an attractive basis for the development of psycholinguistic processing models and practical parsers alike.

For German, most transformational lingusitic theories such as GB posit center-embedding as the underlying word order of sentences with embedded clauses: Weft ich [das Fahrrad zu reparieren] versprochen habe Because I the bike (ace) to repair promised have Because I promised to repair the bike However, far more common is a construction in which the entire subordinate clause is extraposed: Weil ich ti versprochen habe, [das Fahrrad zu reparieren]i.In addition, a third construction is possible, which has been called the "third construction", in which only the embedded verb, but not its nominal argument has been extraposed: Weil ich das Fahrrad ti versprochen habe [zu reparieren]i, A similar construction can also be observed ff there are two levels of embedding.In this case, the number of possible word orders increases from 3 to 30, 6 of which are shown in Figure 1.Of the 30 sentences, 7 are clearly ungrammatical (marked "*"), and 3 are extremely marginal, but not "flat out" (marked "?*").The remaining 20 are acceptable to a greater or lesser degree (marked "ok" or "?").No attempt has been made in the linguistic or computational literature to account for this full range of data.