Contributions in Merlin

null

In this report we present a Prolog tool for the development and testing of natural language grammars called GTU (German: Grammatik-Testumgebung; grammar test environment). GTU offers a window-oriented user interface that allows the development and testing of natural language grammars under three formalisms. In particular it contains a collection of German test sentences and two types of German lexicons. Both of the lexicons can be adapted to a given grammar via an integrated lexicon interface. GTU has been implemented in Prolog both under DOS and UNIX. It was originally developed as a tutoring tool to support university courses on syntax analysis but in its UNIX-version it allows for the development of large grammars.

In this paper we describe our experiences with a tool for the development and testing of natural language grammars called GTU (German: Grammatik-Testumgebumg; grammar test environment). GTU supports four grammar formalisms under a window-oriented user interface. Additionally, it contains a set of German test sentences covering various syntactic phenomena as well as three types of German lexicons that can be attached to a grammar via an integrated lexicon interface. What follows is a description of the experiences we gained when we used GTU as a tutoring tool for students and as an experimental tool for CL researchers. From these we will derive the features necessary for a future grammar workbench.

In computational linguistics, we are often interested in developing grammar formalisms declaratively. However, tractability often becomes a problem then. Therefore, we want to argue for the use of constraint logic programming (CLP), and it is yet interesting to note that most logic based natural language systems have not attempted to employ CLP. Our framework and the prototype system UBS combines logic programming with constraint domains (e.g. typed feature structures and nite sets) and constraint techniques (e.g. coroutining).

Kommerziell verfügbare Maschinelle Übersetzungssysteme können auf den ersten Blick eine erstaunliche Vielzahl syntaktischer Konstruktionen verarbeiten. Erst eine detaillierte Analyse zeigt die spezifischen Defizite auf. Wir haben deshalb eine spezielle Testsuite mit 384 Sätzen (226 EN -> DE, 158 DE -> EN) zusammengestellt, die jeweils unterschiedliche Nebensätze enthalten. Dazu gehören indirekte Aussagesätze und Fragesätze, Adverbialsätze, Relativsätze, sowie Infinitiv-, Partizipial- und Gerundium-Konstruktionen. Nebensätze eignen sich für eine solche Untersuchung besonders gut, weil bei ihrer Übersetzung zahlreiche syntaktische Faktoren eine Rolle spielen. Dazu gehören: * das Problem der Abgrenzung einer Konstituente * das Problem der Funktionsbestimmung des Nebensatzes * das Problem des syntaktischen Transfers in der Übersetzung Da die Struktur und Funktion von Nebensätzen im Deutschen anhand äusserlicher Merkmale (Satzzeichen, Konjunktionen) deutlicher erkennbar ist als im Englischen, treten Probleme vor allem bei der Übersetzung vom Englischen ins Deutsche auf. Unsere Untersuchung behandelt deshalb vor allem diese Übersetzungsrichtung. Untersucht wurden die PC-Systeme Langenscheidts T1 (GMS), Personal Translator Plus (IBM, von Rheinbaben & Busch), Power Translator (Globalink) und Systran (MySoft). Nur in wenigen Fällen scheinen bestimmte Konstruktionen allen Systemen gänzlich unbekannt zu sein (z.B. englische Partizipial-Nebensätze). Bei den anderen Fällen zeigt mindestens ein System, dass es prinzipiell möglich ist, dieses Phänomen korrekt zu behandeln. Die meisten Übersetzungsfehler beruhen auf fehlerhafter Abgrenzung des Nebensatzes vom Hauptsatz, Verwechslung der Nebensatztypen, fehlende semantische Analyse und im Bereich der Synthesefehler die fehlerhafte Wortstellung in der Zielsprache. Kurz zusammengefasst erbrachte die Untersuchung die folgenden Ergebnisse: 1. Die schwierigsten Nebensatzkonstruktionen für die Übersetzungssysteme sind Infinitivkonstruktionen, partizipiale Adverbialsätze und Gerundien. Relativsätze werden gut übersetzt, auch wenn das Relativpronomen fehlt. 2. Von den untersuchten Systemen beherrscht Personal Translator Plus die meisten Nebensatzkonstruktionen. Langenscheidts T1 arbeitet sehr uneinheitlich, manchmal erstaunlich gut und manchmal vollkommen falsch. 3. Die Resultate der Nebensatz-Übersetzungen vom Deutschen ins Englische sind tendenziell besser als in der umgekehrten Richtung.

null

We present real time survey evidence from the UK, US and Germany showing that the labor market impacts of COVID-19 differ considerably across countries. Employees in Germany, which has a well-established short-time work scheme, are substantially less likely to be affected by the crisis. Within countries, the impacts are highly unequal and exacerbate existing inequalities. Workers in alternative work arrangements and in occupations in which only a small share of tasks can be done from home are more likely to have reduced their hours, lost their jobs and suffered falls in earnings. Less educated workers and women are more affected by the crisis.

We show that the Diamond and Mirrlees (1971) linear tax model contains the Mirrlees (1971) nonlinear tax model as a special case. In this sense, the Mirrlees model is an application of Diamond-Mirrlees. We also provide a simple derivation of the Mirrleesian optimal income tax formula from the Diamond-Mirrlees commodity tax formula. In the Mirrlees model, the relevant compensated cross-price elasticities are zero, providing a situation where an inverse elasticity rule holds. We provide four extensions that illustrate the power and ease of our approach, based on Diamond-Mirrlees, to study nonlinear taxation. First, we consider annual taxation in a lifecycle context. Second, we include human capital investments. Third, we incorporate more general forms of heterogeneity into the basic Mirrlees model. Fourth, we consider an extensive margin labor force participation decision, alongside the intensive margin choice. In all these cases, the relevant optimality condition is easily obtained as a direct application of the general Diamond-Mirrlees linear tax formula.

Since information asymmetries have been identified as an important source of bank profits, it may seem that the establishment of information sharing will lead to lower investment in acquiring information. However, banks base their decisions on both hard and soft information, and it is only the former type of data that can be communicated credibly. We show that when hard information is shared, banks will invest more in soft, relationship-specific information. These will lead to more accurate lending decisions, favor small, informationally opaque borrowers, and increase welfare. Since relationship banking focuses on the usage of soft information, the model implies that investment in relationship banking will increase. We test our theory using a large sample of firm-level data from 24 countries.

Traditionally, the virtue of democratic elections has been seen in their role as means of screening and sanctioning shirking public officials. This paper proposes a novel rationale for elections and political campaigns considering that candidates incur psychological costs of lying, in particular from breaking campaign promises. These non-pecuniary costs imply that campaigns influence subsequent behavior, even in the absence of reputational or image concerns. Our lab experiments reveal that promises are more than cheap talk. They influence the behavior of both voters and their representatives. We observe that the electorate is better off when their leaders are elected democratically rather than being appointed exogenously - but only in the presence of electoral campaigns. In addition, we find that representatives are more likely to serve the public interest when their approval rates are high. Altogether, our results suggest that elections and campaigns confer important benefits beyond their screening and sanctioning functions.