Refine
Year of publication
Document Type
- Working Paper (10)
- Article (8)
- Conference Proceeding (3)
Has Fulltext
- yes (21)
Is part of the Bibliography
- no (21)
Keywords
- Lambda-Kalkül (8)
- Operationale Semantik (6)
- Programmiersprache (5)
- Nebenläufigkeit (3)
- Funktionale Programmiersprache (2)
- Pufferspeicher (2)
- Alice ML (1)
- Contextual Equivalence (1)
- Futures (1)
- Glioblastoma (1)
Institute
- Informatik (10)
- Neuere Philologien (10)
- Medizin (1)
Motivated by the question of correctness of a specific implementation of concurrent buffers in the lambda calculus with futures underlying Alice ML, we prove that concurrent buffers and handled futures can correctly encode each other. Correctness means that our encodings preserve and reflect the observations of may- and must-convergence. This also shows correctness wrt. program semantics, since the encodings are adequate translations wrt. contextual semantics. While these translations encode blocking into queuing and waiting, we also provide an adequate encoding of buffers in a calculus without handles, which is more low-level and uses busy-waiting instead of blocking. Furthermore we demonstrate that our correctness concept applies to the whole compilation process from high-level to low-level concurrent languages, by translating the calculus with buffers, handled futures and data constructors into a small core language without those constructs.
Motivated by our experience in analyzing properties of translations between programming languages with observational semantics, this paper clarifies the notions, the relevant questions, and the methods, constructs a general framework, and provides several tools for proving various correctness properties of translations like adequacy and full abstractness. The presented framework can directly be applied to the observational equivalences derived from the operational semantics of programming calculi, and also to other situations, and thus has a wide range of applications.
Various concurrency primitives have been added to sequential programming languages, in order to turn them concurrent. Prominent examples are concurrent buffers for Haskell, channels in Concurrent ML, joins in JoCaml, and handled futures in Alice ML. Even though one might conjecture that all these primitives provide the same expressiveness, proving this equivalence is an open challenge in the area of program semantics. In this paper, we establish a first instance of this conjecture. We show that concurrent buffers can be encoded in the lambda calculus with futures underlying Alice ML. Our correctness proof results from a systematic method, based on observational semantics with respect to may and must convergence.
We investigate methods and tools for analyzing translations between programming languages with respect to observational semantics. The behavior of programs is observed in terms of may- and mustconvergence in arbitrary contexts, and adequacy of translations, i.e., the reflection of program equivalence, is taken to be the fundamental correctness condition. For compositional translations we propose a notion of convergence equivalence as a means for proving adequacy. This technique avoids explicit reasoning about contexts, and is able to deal with the subtle role of typing in implementations of language extensions.
We investigate methods and tools for analyzing translations between programming languages with respect to observational semantics. The behavior of programs is observed in terms of may- and mustconvergence in arbitrary contexts, and adequacy of translations, i.e., the reflection of program equivalence, is taken to be the fundamental correctness condition. For compositional translations we propose a notion of convergence equivalence as a means for proving adequacy. This technique avoids explicit reasoning about contexts, and is able to deal with the subtle role of typing in implementations of language extensions.
We investigate methods and tools for analysing translations between programming languages with respect to observational semantics. The behaviour of programs is observed in terms of may- and mustconvergence in arbitrary contexts, and adequacy of translations, i.e., the reflection of program equivalence, is taken to be the fundamental correctness condition. For compositional translations we propose a notion of convergence equivalence as a means for proving adequacy. This technique avoids explicit reasoning about contexts, and is able to deal with the subtle role of typing in implementations of language extensions.
We investigate methods and tools for analysing translations between programming languages with respect to observational semantics. The behaviour of programs is observed in terms of may- and mustconvergence in arbitrary contexts, and adequacy of translations, i.e., the reflection of program equivalence, is taken to be the fundamental correctness condition. For compositional translations we propose a notion of convergence equivalence as a means for proving adequacy. This technique avoids explicit reasoning about contexts, and is able to deal with the subtle role of typing in implementations of language extensions.
Motivated by the question of correctness of a specific implementation of concurrent buffers in the lambda calculus with futures underlying Alice ML, we prove that concurrent buffers and handled futures can correctly encode each other. Correctness means that our encodings preserve and reflect the observations of may- and must-convergence, and as a consequence also yields soundness of the encodings with respect to a contextually defined notion of program equivalence. While these translations encode blocking into queuing and waiting, we also describe an adequate encoding of buffers in a calculus without handles, which is more low-level and uses busy-waiting instead of blocking. Furthermore we demonstrate that our correctness concept applies to the whole compilation process from high-level to low-level concurrent languages, by translating the calculus with buffers, handled futures and data constructors into a small core language without those constructs.
Purpose: The role of obesity in glioblastoma remains unclear, as previous analyses have reported contradicting results. Here, we evaluate the prognostic impact of obesity in two trial populations; CeTeG/NOA-09 (n = 129) for MGMT methylated glioblastoma patients comparing temozolomide (TMZ) to lomustine/TMZ, and GLARIUS (n = 170) for MGMT unmethylated glioblastoma patients comparing TMZ to bevacizumab/irinotecan, both in addition to surgery and radiotherapy.
Methods: The impact of obesity (BMI ≥ 30 kg/m2) on overall survival (OS) and progression-free survival (PFS) was investigated with Kaplan–Meier analysis and log-rank tests. A multivariable Cox regression analysis was performed including known prognostic factors as covariables.
Results: Overall, 22.6% of patients (67 of 297) were obese. Obesity was associated with shorter survival in patients with MGMT methylated glioblastoma (median OS 22.9 (95% CI 17.7–30.8) vs. 43.2 (32.5–54.4) months for obese and non-obese patients respectively, p = 0.001), but not in MGMT unmethylated glioblastoma (median OS 17.1 (15.8–18.9) vs 17.6 (14.7–20.8) months, p = 0.26). The prognostic impact of obesity in MGMT methylated glioblastoma was confirmed in a multivariable Cox regression (adjusted odds ratio: 2.57 (95% CI 1.53–4.31), p < 0.001) adjusted for age, sex, extent of resection, baseline steroids, Karnofsky performance score, and treatment arm.
Conclusion: Obesity was associated with shorter survival in MGMT methylated, but not in MGMT unmethylated glioblastoma patients.
We investigate methods and tools for analysing translations between programming languages with respect to observational semantics. The behaviour of programs is observed in terms of may- and must-convergence in arbitrary contexts, and adequacy of translations, i.e., the reflection of program equivalence, is taken to be the fundamental correctness condition. For compositional translations we propose a notion of convergence equivalence as a means for proving adequacy. This technique avoids explicit reasoning about contexts, and is able to deal with the subtle role of typing in implementations of language extension.
This paper argues that short (clause-internal) scrambling to a pre-subject position has A properties in Japanese but A'-properties in German, while long scrambling (scrambling across sentence boundaries) from finite clauses, which is possible in Japanese but not in German, has A'-properties throughout. It is shown that these differences between German and Japanese can be traced back to parametric variation of phrase structure and the parameterized properties of functional heads. Due to the properties of Agreement, sentences in Japanese may contain multiple (Agro- and Agrs-) specifiers whereas German does not allow for this. In Japanese, a scrambled element may be located in a Spec AgrP, i.e. an A- or L-related position, whereas scrambled NPs in German can only appear in an AgrP-adjoined (broadly-L-related) position, which only has A'-properties. Given our assumption that successive cyclic adjunction is generally impossible, elements in German may not be long scrambled because a scrambled element that is moved to an adjunction site inside an embedded clause may not move further. In Japanese, long distance scrambling out of finite CPs is possible since scrambling may proceed in a successive cyclic manner via embedded Spec- (AgrP) positions. Our analysis of the differences between German and Japanese scrambling provides us with an account of further contrasts between the two languages such as the existence of surprising asymmetries between German and Japanese remnant-movement phenomena, and the fact that unlike German, Japanese freely allows wh-scrambling. Investigation of the properties of Japanese wh-movement also leads us to the formulation of the "Wh-cluster Hypothesis", which implies that Japanese is an LF multiple wh-fronting language.
Expletives as features
(2000)
Expletives have always been a central topic of theoretical debate and subject to different analyses within the different stages of the Principles and Parameter theory (see Chomsky 1981, 1986, 1995; Lasnik 1992, 1995; Frampton and Gutman 1997; among others). However, most analyses center on the question how to explain the behavior of expletives in A-chains (such as there in English or Þad in Icelandic). No account relates wh-expletives (as one finds them in so-called partial wh-movement constructions in languages such as Hungarian, Romani, and German) to expletives in Achains. In this paper, I argue that the framework of the Minimalist Program opens up the possibility of accounting for expletive-associate relations in A-/A'-chains in a unified manner. The main idea of the unitary analysis is that an expletive is an overtly realized feature bundle that is (sub)extracted from its associate DP. There in an expletive-associate chain is a moved D-feature which orginates inside the associate DP. Similarily, in A'-chains, the whexpletive originates as a focus-/wh-feature in the wh-phrase with which it is associated. This analysis provides evidence for the feature-checking theory in Chomsky (1995). The paper is organized as follows. Section 2 contains the discussion of expletive there. In section 3 I suggest an analysis for whexpletives, and I also explore whether this analysis can be extended to relations between X°-categories such as auxiliary and participle complexes.
In this article, I discuss some important properties of wh-questions and wh-scrambling in Japanese. The questions I will address are (i) which instances of (wh-) scrambling involve reconstruction and (ii) how the undoing effects of scrambling can be derived. First I will discuss the claim that (wh-) scrambling is semantically vacuous and is therefore undone at LF (Saito 1989, 1992). Then I consider the data that led Takahashi (1993) to the conclusion that at least some instances of wh-scrambling have to be analyzed as instances of "full wh-movement" i.e., overt movement of the wh-phrase in its scopal position. It will be argued that these examples are not instances of full wh-movement in Japanese, but that they also represent semantically vacuous scrambling. Those instances of scrambling that apprently cannot be undone are best explained with recourse to parsing effects. I conclude that wh-scrambling in Japanese is always triggered by a ([-wh]-) scrambling feature. In addition, long distance scrambling (scrambling out of finite CPs) is analyzed as adjunction movement, whereas short distance scrambling is movement to a specifier position of IP. Turning to the mechanisms of undoing, I will argue that only long distance scrambling is undone. This is shown to follow from Chomsky's (1995) bare phrase structure analysis, according to which multi-segmental categories derived by adjunction movement are not licensed at LF. The article is organized as follows. In section 2, the wh-scrambling phenomenon is described. In section 3, I discuss the reconstruction properties of scrambling. In addition, this section provides some basic assumptions about my analysis of Japanese scrambling in general. In section 4, I turn to the analysis of wh-scrambling as an instance of full wh-movement in Japanese. Section 5 provides discussion of multiple wh-questions in Japanese, and section 6 gives the conclusion.
The languages of the world differ with respect to argument extraction possibilities. In languages such as English, wh-movement is possible from Spec IP and from the complement position, whereas in languages such as Malagasy only extraction from Spec IP is possible. This difference correlates with the fact that these language types obey different island constraints and behave differently with respect to wh-in situ and superiority effects. The goal of this paper is to outline an analysis for these differences. The basic idea is that in contrast to languages such as English, in Malagasy-type languages every argument can be merged in the complement position of the selecting head.
Die Doppelobjekt-Konstruktion bildet einen Untersuchungsgegenstand, der in der Vergangenheit die Theoriebildung in der Syntaxforschung wesentlich beeinflusst hat. Untersuchungen zu Doppelobjekt-Konstruktionen sind u.a. folgenreich gewesen für die Kasustheorie sowie für Analysen der Verbbewegung, Satz-, VP- und Argument-Struktur. In diesem Aufsatz stelle ich eine Analyse einiger wichtiger Aspekte der Doppelobjekt-Konstruktion im Deutschen vor. Untersucht wird, in welcher Position die Objekte des Verbs basisgeneriert werden und in welchen abgeleiteten Positionen sie erscheinen. Die Beantwortung dieser Fragen liefert eine Erklärung für das asymmetrische Verhalten der beteiligten Objekte in Bezug auf ihr Bindungs- und Extraktionsverhalten.
In diesem Aufsatz gehe ich der Frage nach, in wie viel unterschiedlichen Positionen Verben im deutschen Satz vorkommen können. Anhand syntaktischer Tests wird gezeigt, daß das Verb im Deutschen in insgesamt drei unterschiedlichen Positionen auftritt und nicht, wie in der traditionellen Grammatik angenommen wird, in nur zwei Positionen (in der rechten und in der linken Satzklammer). Es wird dafür argumentiert, daß die Anwendung des abstrakten Satzschemas, wie es heute gängigerweise in der generativen Grammatikforschung als universelles Satzmodell angenommen wird, die Erklärung einer Vielzahl syntaktischer Phänomene im Deutschen ermöglicht, die mit der traditionellen Verbstellungsanalyse, die von nur zwei Verbpositionen ausgeht, nicht erklärt werden können. Gemäß des universellen Satzschemas repräsentiert die Infl(ection)-Position eine Verbposition im Satz. Diese Position ist identisch mit der rechten Satzklammer. Eine weitere Verbposition ist die V-Position innerhalb des Mittelfelds und eine dritte potenzielle Position für das Verb entspricht der C(omplementizer-) Position (bzw. der linken Satzklammer). Der Aufsatz ist folgendermaßen gegliedert. In der Einleitung schildere ich kurz die unterschiedlichen Auffassungen, die in der Vergangenheit zur Verbstellungsproblematik im Deutschen vertreten wurden. In Abschnitt 2 nenne ich die wichtigsten Argumente, die gegen die Annahme vorgebracht wurden, daß im deutschen Satz für Verben insgesamt drei Positionen zur Verfügung stehen. Im Anschluss daran werden in den Abschnitten 3.1 bis 3.2 Argumente diskutiert, die für drei Verbpositionen sprechen. Abschnitt 3.3 behandelt die Frage, wie vor diesem Hintergrund die Daten aus Abschnitt 2, die sich als problematisch für diese Analyse erwiesen haben, erklärt werden können. In Abschnitt 4 wende ich mich weiterer unabhängiger Evidenz aus dem Bereich der historischen Syntax zu, die für drei Verbpositionen im Deutschen spricht. In Abschnitt 5 gebe ich eine kurze Zusammenfassung der wichtigsten Ergebnisse.
In diesem Aufsatz diskutiere ich die Distribution von kohärenten Kontroll-Infinitiven im Deutschen. Es werden die Verbklassen bestimmt, die kohärente Infinitive lizenzieren. Dabei zeigt sich, dass ausschließlich Infinitive 'kohärent' konstruiert werden können, die die Position der Akkusativ NP (bzw. die Position des direkten Objekts) einnehmen. Kontroll-Infinitive in anderen strukturellen Positionen sind zwangsläufig 'inkohärent'. Transparente Infinitive in Sprachen wie dem Polnischen und Spanischen sind in derselben Weise in ihrer Distribution beschränkt. Ich schlage eine einheitliche Analyse der relevanten Daten vor, die weitere distributionelle Generalisierungen bezüglich des Auftretens kohärenter Infinitive korrekt prognostiziert. Für die idiolektale Variation, die unter Sprechern in Bezug auf bestimmte Verben existiert, die kohärente Infinitive lizenzieren, wird eine Erklärung formuliert, die auf der Idee basiert, dass die Bildung dieser Infinitive an die Präsenz eines Inkorporations-Merkmals gebunden ist, das beim Spracherwerb auf der Grundlage positiver Evidenz erworben wird. In this article, I discuss the distribution of so-called 'coherent (control) infinitives' in German. In section 2, I will argue that both coherent as well as incoherent control-infinitives have a sentential status. In section 3, I argue that only infinitives occupying the position of the direct object show the well-known properties associated with coherent infinitives. Control infinitives in other structural positions represent incoherent infinitives. This situation is not limited to German infinitives. Transparent infinitives in Polish and Spanish show the same structural asymmetry. In section 4, I propose a unified analysis for the data that in addition, correctly predicts further restrictions for the distribution of coherent infinitives. In section 5, I propose an account for idiolectal variation in the class of verbs that license coherent infinitives. This account is based on the idea that coherent infinitives require an incorporation-feature in their lexical entry that is acquired on the basis of positive evidence.
In this paper I show that Clitic Climbing (CC) in Spanish and Long Scrambling (LS) in German (and Polish) are (im-)possible out of the same environments. For an explanation of this fact I propose a feature-oriented analysis of incorporation phenomena. The idea is that restructuring is a phenomenon of syntactic incorporation. In German and Polish, Agro incorporates covertly into the matrix clause and licenses LS out of the infinitival into the matrix clause. Similarily the clitic in Spanish, which is analysed as an Agro-head, incorporates into the matrix clause. I argue that this movement is necessary for reasons of feature-checking, i. e. for checking of an [+R]- or Restructuring-feature. In section 2 I discuss several differences between CC and LS. For example, the proposed analysis correctly predicts that clitics in contrast to scrambled phrases are subject to several serialization restrictions. Throughout the paper I use the term restructuring only in a descriptive sense, in order to describe the phenomenon in question.
Ausgangspunkt der folgenden Untersuchung ist die Überlegung, daß verschiedene Versionen der Prinzipien- und Parametertheorie unterschiedliche Prognosen bezüglich strukturell ambiger Wortfolgen in Passiv-Konstruktionen des Deutschen machen. Im Rahmen einer Theorie, in der Move-alpha frei appliziert, wie etwa in der Rektions- und Bindungstheorie (Chomsky 1981, 1986a, 1986b), können multiple Derivationen für derartige Abfolgen nicht ausgeschlossen werden, wohingegen eine andere Situation vorliegt, wenn man die entsprechenden Konstruktionen im Rahmen des Minimalistischen Programms analysiert. Hier kann die Anzahl möglicher (und mit einer Wortfolge verträglicher) Derivationen mit Hilfe von Ökonomieprinzipien beschränkt werden. Auf Grundlage verschiedener syntaktischer Tests wird im weiteren gezeigt, daß bestimmte Wortfolgen nur mit einer Derivation verträglich sind, was im Einklang mit einer minimalistischen Analyse der Daten steht. Der Aufsatz ist folgendermaßen gegliedert. In Abschnitt 2 erläutere ich das Grundproblem der multiplen Derivationen, das sich im Deutschen z. B. bei Passiv-Konstruktionen ergibt, wenn man annimmt, daß NP-Bewegung und Scrambling optional erfolgen und ferner keine Beschränkungen für potentielle Derivationen gelten. In Abschnitt 3 diskutiere ich die Voraussetzungen für die Überprüfung der Prognosen der verschiedenen Varianten des Prinzipien- und Parametermodells und versuche anschließend auf Grundlage syntaktischer Tests zu belegen, daß die diskutierten Beispiele tatsächlich nicht strukturell ambig, sondern strukturell eindeutig sind, wie es die Analyse der entsprechenden Konstruktionen im Rahmen des Minimalistischen Programms vorhersagt. Abschnitt 4 beschreibt die Konsequenzen der Analyse für weitere Sprachen wie Niederländisch und Japanisch und zusätzliche Bewegungstypen. Abschnitt 5 enthält die Konklusion.