Ordinal Rankings on Measures Annihilating Thin Sets

نویسندگان

  • Alexander S. Kechris
  • Russell Lyons
  • ALEXANDER S. KECHRIS
  • RUSSELL LYONS
چکیده

We assign a countable ordinal number to each probability measure which annihilates all H-sets. The descriptive-set theoretic structure of this assignment allows us to show that this class of measures is coanalytic non-Borel. In addition, it allows us to quantify the failure of Rajchman's conjecture. Similar results are obtained for measures annihilating Dirichlet sets. A closet subset E of the unit circle T R/Z is called an H-set if there exists a sequence {nk} of positive integers tending to oo and an interval (i.e., a nonempty open arc) I C T such that for all k and all x E E, nkx 0 I. These sets play a fundamental role as examples of sets of uniqueness for trigonometric series [KL; Z, Chapters IX, XII]. A (Borel) probability measure ,u on T is called a Rajchman measure if ,(n) -+ 0 as Il -+ oo, where ,u(n) = fTTe(-nx)dp(x), e(x) = e2rix. We denote by R the class of such measures. These measures have also been very important o the study of sets of uniqueness. In particular, every Rajchman measure annihilates every set of uniqueness, hence every H-set. After establishing these relationships [Rl, R2], Rajchman conjectured that, in fact, the only measures which annihilate all H-sets are those in R. This, however, is false [Li, L2, L3, L5]. Here, we shall quantify how distant Rajchman's conjecture is from the truth. Given a class ' of closed subsets of T, denote by WJ" the class of probability measures on T which annihilate all sets in i?: , E F" X* VE E ' (,(E) = 0). Thus R C H', where H denotes the class of H-sets. Denote by PROB(T) the compact, metrizable space of (Borel) probability measures on T with the weak* topology. It is easy to check that R is a Borel, in fact II3 (i.e., F6), subset of this space. We establish in ?3 that H' is a rI1 (i.e., coanalytic) but not Borel subset of PROB(T). This is the first example of a natural class of measures of such complexity known to the authors and it highlights the distinction between R and H'. Our method of proof actually provides quite a bit of further information on the relationship between R and H'. In ?1, we assign to each ,u E H' a countable ordinal number h(s) which measures in some sense the complexity of the verification that ,u annihilates all H-sets. We show that h has certain definability properties, namely, it is a Il1-rank (see [KL]). In ?3, using the techniques developed in [L3 and L4], we prove that the rank h is unbounded in w1, the first uncountable ordinal; Received by the editors September 10, 1987. 1980 Mathematics Subject Classification (1985 Revision). Primary 43A46, 03E15; Secondary 43A05, 42A63, 54H05.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Ranking Law Journals and the Limits of Journal Citation Reports

Rankings of schools, scholars, and journals emphasize ordinal rank. Journal rankings published by Journal Citation Reports (JCR) are widely used to assess research quality, which influences important decisions by academic departments, universities, and countries. We study refereed law journal rankings by JCR, Washington and Lee Law Library (W&L), and the Australian Research Council (ARC). Both ...

متن کامل

Using and Learning GAI-Decompositions for Representing Ordinal Rankings

We study the use of GAI-decomposable utility functions for representing ordinal rankings on combinatorial sets of objects. Considering only the relative order of objects leaves a lot of freedom for choosing a particular utility function, which allows one to get more compact representations. We focus on the problem of learning such representations, and give a polynomial PAC-learner for the case ...

متن کامل

Subjective Performance Evaluations and Employee Careers

Subjective Performance Evaluations and Employee Careers Firms commonly use supervisor ratings to evaluate employees when objective performance measures are unavailable. Supervisor ratings are subjective and data containing supervisor ratings typically stem from individual firm level data sets. For both these reasons, doubts persist on how useful such data are for evaluating theories in personne...

متن کامل

Democracy and Social Choice: A Response to Saari

1 Ethics 111 (2001): 706-34; I refer to that article as “my earlier article.” This article is a response to Donald G. Saari (2003), “Capturing the ‘Will of the People’,” Ethics 113 (2003): 333-349. Thanks to Hélène Landemore and Richard Zeckhauser for valuable comments on an earlier draft. I write (A, B, etc.) for rankings, and {A, B, etc.} for sets. Where Saari’s notation differs, I substitute...

متن کامل

Ordinal and cardinal measures of health inequality: an empirical comparison.

When measuring health inequality using ordinal data, analysts typically must choose between indices specifically based upon ordinal data and more standard indices using ordinal data, which has been transformed into cardinal data. This paper compares inequality rankings across a number of different approaches and finds considerable sensitivity to the choice between ordinal- and cardinal-based in...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2010