I 39 ~
Law/invent/discover/Goodman: to the discovery of laws, belongs to design them.
>
Discoveries.
---
II 37
E.g. ... because he describes a random fact and there is no law.
And apparently also no purely syntactic criterion can be useful, because the most specific description of individual facts can be brought into a form that has any desired degree of syntactic generality.
II 38
I just want to emphasize the thought by Hume that a sentence is not used for predictions because it is a law, but that it is called a law, because it is used for predictions. And that the law is not used for predictions, because it describes a causal connection, but that the meaning of the causal connection is to explain by the help of laws for predictions.
>
Predictions.
II 40f
Definition Act (wrong): a law is a sentence that is lawlike and true - but a sentence can be true but not lawlike, or lawlike and not true.
For this definition lawlikeness would be a short-lived and random affair. Only sentences that you actually use for predictions, would be lawlike. And a true sentence which has been used for predictions, would be no law anymore if it had once been fully examined.
II 41
Lawlike/Goodman: a sentence is lawlike if its recognition does not depend on the decision of any given application case alone.
II 41
Sensible is that there schould be no application case on which test the recognition depends on. This criterion does not allow statements like "This book is black and oranges are round" to be lawlike because their recognition is subject to the knowledge, if this book is black.
II 109
Lawlike or resumable hypotheses are not to be characterized in a purely syntactic way.
>
Hypotheses.
II 114
If all application cases are examined, there is no hypothesis or law anymore.
II 114
The hypothesis neither needs to be true nor false, nor lawlike or even just reasonable, because we do not speak of what should be continued, but what is actually continued. Wrong hypotheses can be supported.
