The archive.

Distinguishing two concepts of probability need not mean separating them. In this brief oral reply to Professor Bartlett, Gerhard Tintner argues that discussions of probability often founder on conceptual confusion rather than genuine disagreement. He connects degree of confirmation with relative frequency: as evidence accumulates, the former must approach the latter. Yet convergence does not settle how to assign measures when evidence is scarce. This two-paragraph contribution isolates the point at which Tintner sees Carnap’s theory as useful—the small sample, where initial assignments still matter—and clarifies why a distinction between concepts can support, rather than obstruct, an account of their relationship.

Can a logical measure of confirmation genuinely learn from experience if its starting weights are chosen for simplicity? In this brief written reply to a journal discussion, Gerhard Tintner defends Carnap’s inductive logic on precisely that ground. He contrasts a weighting procedure that leaves confirmation untouched by previous observations with a predictive rule whose results approach observed relative frequencies as samples grow. His defence remains qualified: the language must satisfy demanding conditions, the notion of a simple predicate presents difficulties, and application to continuous variables awaits further theory. The reply offers a compact view of how Tintner tests a formal construction against empirical learning without concealing its unfinished foundations.

How much support does evidence give a scientific theory, apart from the costs of acting on it? Gerhard Tintner approaches this question from econometrics, where choosing between rival economic theories cannot readily be reduced to a calculation of gains and losses. In this 1949 article, he develops Rudolf Carnap’s logical account of confirmation through worked statistical examples. Coin tosses expose a consequential choice: assigning equal weight to patterns of outcomes rather than to individual sequences changes the probabilities used for prediction. Readers can trace how the language used to describe evidence enters the resulting estimates, and why confirming a universal law differs from predicting its next instance. Tintner also marks the framework’s limits: its treatment of categorical attributes does not yet accommodate the continuous variables central to economics and physics.

What can economic measurement establish—and what must remain beyond its authority? In this 1949 French article, Gerhard Tintner places econometrics between mathematical deduction and statistical description: theory makes assumptions explicit, while observations test their consequences. His distinctive concern is how to identify economically meaningful relationships in data that do not come from controlled experiments. An estimate of agricultural demand shows what numerical analysis can reveal; a discussion of tariffs shows why estimating consequences cannot decide which social goals to pursue. Readers encounter a defence of econometrics that also marks its limits, distinguishing the testing of individual coefficients from the still unresolved task of judging competing economic theories as wholes.

Toutes les lois économiques sont conditionnelles.

English translation: “All economic laws are conditional.”

What can an estimated demand curve tell policymakers—and what can it never decide for them? In this 1949 article, Gerhard Tintner connects the methodological claims of econometrics to the difficulties of measuring American agricultural demand and supply. Prices and quantities do not identify those relationships unaided: economic assumptions must first make them distinguishable. His empirical results sharpen the point. Demand estimates yield interpretable elasticities, while insignificant supply coefficients expose missing influences, including weather and possibly lagged prices. Tintner treats such failures as grounds for revising models, not concealing uncertainty. The article offers a concrete account of how theory becomes a testable numerical claim, while keeping estimates of policy consequences distinct from judgements about which social objectives deserve priority.

Do British industrial employment and labor supply respond to real wages, or separately to money wages and prices? In this brief conference-paper abstract, Gerhard Tintner reports estimates for 1920–1938, distinguishing employers’ wholesale-price measure from workers’ cost of living. His demand estimates suggest that the wage–price ratio matters, but the estimated elasticity changes from −0.4 to −0.7 when a time trend is included. On supply, he is more guarded: the negative elasticities are interesting, yet dependence on wages and prices remains uncertain. The abstract offers a compact encounter with an empirical tension—how much an econometric estimate can establish when its magnitude depends on specification and the underlying relationship itself is doubtful.

Different statistical aims can lead to closely related estimating equations without making the methods interchangeable. In this mathematical article, Gerhard Tintner connects canonical correlation, principal components, weighted regression, and discriminant analysis through constrained optimization. His distinctive move is to identify the matrices and normalization conditions that give a shared algebraic structure its different statistical meanings: a covariance matrix paired with the identity in principal components, for example, becomes sample covariance paired with error covariance in weighted regression. Readers can discover how apparently separate procedures meet through reductions to common stationary equations—and why their objectives and assumptions still matter. Tintner confines the comparison to estimation, leaving sampling distributions outside the argument.

Gerhard Tintner’s 1950 review of Jan Tinbergen’s The Dynamics of Business Cycles asks what specialized econometric research can offer economists without advanced mathematical training. His answer rests on Tinbergen’s combination of accessible exposition, statistical investigation, and experience in Dutch economic planning. Reviewing the translation adapted for American readers by J. J. Polak, Tintner values a Keynesian-influenced approach that nevertheless considers other theoretical systems. His particular attention to hog and coffee markets gives the review a concrete agricultural emphasis: aggregate fluctuations matter alongside cycles in individual commodities. This short, favorable assessment shows why Tintner regarded Tinbergen’s synthesis as useful for teaching and policy discussion, without offering a detailed critique of its models or prescriptions.

Writing in tribute to Abraham Wald, Gerhard Tintner asks what mathematical rigor can secure for economics—and where its assumptions limit practical use. This 1952 memorial survey distinguishes the formulation of equilibrium equations from proofs that economically admissible solutions exist. It also shows how cost-of-living comparisons depend on information about preferences that observed prices and purchases alone cannot supply. Tintner’s appreciation is not uncritical: he questions minimax decision rules that treat an indifferent Nature as an adversary, and the feasibility of measuring social losses for policy decisions. The article offers a compact encounter with Wald’s achievements through an economist’s discriminating perspective, revealing both the power of explicit assumptions and the empirical work still needed to make formal results useful.

When observations contain measurement error, how can one determine whether their underlying systematic components obey independent linear relations? In this compact 1952 article, Gerhard Tintner makes the error covariance matrix the link between multicollinearity and weighted regression. For time series with smooth systematic components, he proposes estimating error covariances through successive differences, then assessing observed covariance against that noise structure. The same calculation supports both an estimate of the number of underlying relations and the coefficients needed to express them. Readers can examine a precise connection between error estimation and regression, together with its limits: errors must have constant covariances and no lag correlation, while the significance tests rely on normality and large-sample approximations rather than a known exact distribution.

A fitted relationship need not be the economic relationship an investigator seeks. In Econometrics, first published in 1952, Gerhard Tintner makes this gap between statistical calculation and economic interpretation a central teaching problem. Prices and quantities are jointly determined, so reversing a simple regression cannot by itself recover supply or demand; smoothing a time series can introduce dependence rather than merely uncover it. Tintner’s emphasis is statistical, but his tests of method remain tied to economic questions and worked applications. Readers can discover why assumptions about structure, measurement and temporal dependence change what an estimate warrants—and why describing past observations is not enough to justify a forecast. This reprint preserves a textbook concerned as much with the limits of quantitative inference as with its procedures.

What must be assumed before an economic equation can yield a credible number? In this 1953 article, Gerhard Tintner presents econometrics through the gap between theoretical relationships and measurable effects. Demand and supply observations do not automatically disclose separate demand and supply curves; a precise estimate does not establish that its underlying model is adequate. Examples from meat and corn markets make these difficulties concrete, showing how external variables, past prices, and assumptions about error shape what can be inferred. Tintner’s perspective joins technical explanation to methodological caution: numerical estimates can test economic claims and inform choices about taxes or subsidies, but cannot determine policy goals. Readers can discover why the usefulness of an economic coefficient depends on more than its statistical precision.