Left examples have the keyword "teach" on the OEBP.

Sometimes instead of gauging somebody's ability to guess the pattern, a Bongard Problem might teach the pattern.

Consider a Bongard Problem whose left examples are images of a specific person's face; after seeing that Problem, one might be able to recognize that person.

A "teach" Bongard Problem (with a huge number of examples) could be taken as a training set for machine learning.

"Teach" BPs tend to be convoluted, arbitrary, cultural-knowledge-based (keyword culture), or they illustrate some insight that might be overlooked, perhaps mathematical (keyword math).

Adjacent-numbered pages:
BP853 BP854 BP855 BP856 BP857  *  BP859 BP860 BP861 BP862 BP863

meta (see left/right), links, keyword

bp [smaller | same | bigger]

Aaron David Fairbanks

Left-sorted Bongard Problems have the keyword "consistentsymbols" on the OEBP.

A most extreme "consistentsymbols" Bongard Problem is BP121: the solution is about codes consistently symbolizing objects. However, "consistentsymbols" Bongard Problems may have solution unrelated to the symbolism; the symbolism may just be implicit, e.g. always meaning dots as numbers, always meaning stacked dots as fractions, repeatedly using the same simple drawings as shorthand to represent platonic solids. Most BPs have some symbolism in this sense; a Bongard Problem should only be labelled "consistentsymbols" if there is a relatively high amount of varied symbolism, particularly if it is visual symbolism not all people would naturally understand.

A Bongard Problem featuring a real language would be another extreme example of "consistentsymbols".

A Bongard Problem with many varied images meant to be interpreted in unique ways is not necessarily "consistentsymbols," since there is no specific-to-this-Bongard-Problem vocabulary of symbols that must be known to understand it. (Even so, some might say that how people intuitively interpret images is a vocabulary on its own.)

Sometimes, the symbolism isn't an important part of the Bongard Problem, and it just helps make the Bongard Problem easier to read (see the help keyword). For example, a Bongard Problem may include many clumps of dots, and the solution of the Problem may have to do with counting the number of dots in each clump; the Bongard Problem might build up a symbolic context by always arranging each number of dots in a consistent way (e.g. how they conventionally appear on dice faces).

"Consistentsymbols" is related to the keyword structure, a format that all examples fit that the solver needs to know how to read. In "consistentsymbols" Bongard Problems, not all examples need to fit a rigid format; instead there may be various smaller structures of meaning that only appear in some examples.

"Consistentsymbols" is related to assumesfamiliarity, BPs that require the solver to take certain assumptions about what the examples are for the solution to seem simple. A "consistentsymbols" Bongard Problem may have a very convoluted solution that involves explaining the meaning of each appearing object; however, the solution can become simple given correct interpretations of all objects. This effect works best when each object must be interpreted the same way across all boxes in order for the simple solution to fit. The comments sections of "consistentsymbols" BP pages on the OEBP ought to explain the symbolism used.

Adjacent-numbered pages:
BP833 BP834 BP835 BP836 BP837  *  BP839 BP840 BP841 BP842 BP843

meta (see left/right), links, keyword, wellfounded

visualbp [smaller | same | bigger]

Aaron David Fairbanks

Left examples have the keyword "collective" on the OEBP.

Some Bongard Problems are "collective" in a more extreme way than others. Perhaps there are absolutely no individual examples that anyone would confidently sort on either side, and the solver can only be expected to get a vague gist by seeing them all together. Or perhaps in practice most people agree about where most examples should fit, even though a stretch of an argument could conceivably be made for each one fitting on the other side.

In some collective Bongard Problems, each example admits a number of possible interpretations, and the correct choice of interpretation is only clear once the solution is known. The group of examples together improve the solver's confidence about having understood each individual one right. This is common in rules Bongard Problems), where each example communicates its own rule.

Collective Bongard Problems are borderline invalid Bongard Problems (see https://www.oebp.org/invalid.php ). There is no one rule dividing the sides; the solution is not a method to determine whether an arbitrary example fits left or right. It is a less strict kind of Bongard Problem.

Collective implies fuzzy.

Collective Bongard Problems are often abstract".

Subjective Bongard Problems are often collective.

In some Bongard Problems, each example has a corresponding slightly different twin example on the other side (keyword contributepairs), and it is necessary to see both examples together in order to be able to sort either of them. This is related to "collective" but not quite the same. It becomes unambiguous where an example fits once its twin is seen.

Adjacent-numbered pages:
BP832 BP833 BP834 BP835 BP836  *  BP838 BP839 BP840 BP841 BP842

meta (see left/right), links, keyword

bp [smaller | same | bigger]

Aaron David Fairbanks

Every Bongard Problem on the left shows invariance under some transformation in left sub-examples.

Some Bongard Problems on the right are symmetrical or have sub-examples that are symmetrical.

See BP760 for the version with links to pages on the OEBP instead of images of Bongard Problems.

Adjacent-numbered pages:
BP831 BP832 BP833 BP834 BP835  *  BP837 BP838 BP839 BP840 BP841

meta (see left/right), miniproblems, creativeexamples, assumesfamiliarity, structure, presentationinvariant

boxes_bpimage_three_per_side [smaller | same | bigger]

Aaron David Fairbanks

"Tiling" is placing shapes next to each other without overlap to fill up space or other shapes.

See BP706 for the version with links to pages on the OEBP instead of images of Bongard Problems.

Adjacent-numbered pages:
BP830 BP831 BP832 BP833 BP834  *  BP836 BP837 BP838 BP839 BP840

meta (see left/right), miniproblems, creativeexamples, assumesfamiliarity, structure, presentationinvariant

boxes_bpimage_three_per_side [smaller | same | bigger]

Aaron David Fairbanks

When removal of a box changes the solution, it cannot remove the existing solution; it can only allow more possible new solutions. Conversely, adding boxes can only narrow down between existing solutions.

The special box is always placed in the bottom left for help, but another solution is "there is some box whose removal allows another solution vs. not".

Bongard Problems sorted left might be called "deceptive". Especially when the solution ruled out by the one box is the more obvious solution.

Adjacent-numbered pages:
BP829 BP830 BP831 BP832 BP833  *  BP835 BP836 BP837 BP838 BP839

meta (see left/right), miniproblems, creativeexamples, presentationmatters, assumesfamiliarity, structure

boxes_bpimage_three_per_side [smaller | same | bigger]

Aaron David Fairbanks

All examples are "object present vs. object absent" Problems.

Adjacent-numbered pages:
BP828 BP829 BP830 BP831 BP832  *  BP834 BP835 BP836 BP837 BP838

abstract, meta (see left/right), miniproblems, creativeexamples, assumesfamiliarity, structure, presentationinvariant

[smaller | same | bigger]

Aaron David Fairbanks

"Specific object" we mean the same connected figure (up to movement around the box). It is ambiguous whether a scaled, rotated, or reflected version of an object would be considered "the same object" because no examples on either side are forced to deal with that subtlety in their respective solutions.

Adjacent-numbered pages:
BP827 BP828 BP829 BP830 BP831  *  BP833 BP834 BP835 BP836 BP837

meta (see left/right), miniproblems, assumesfamiliarity, structure, presentationinvariant

[smaller | same | bigger]

Aaron David Fairbanks

Right examples are Bongard Problems in which the left hand side is enough to communicate the full answer.

No examples shown have right side with multiple rules and left side narrowing down the answer.

Adjacent-numbered pages:
BP826 BP827 BP828 BP829 BP830  *  BP832 BP833 BP834 BP835 BP836

hard, meta (see left/right), miniproblems, creativeexamples, presentationmatters, assumesfamiliarity, structure, contributepairs

boxes_bpimage_three_per_side [smaller | same | bigger]

Aaron David Fairbanks

Left: the left hand side is enough to communicate the answer; the left pattern can be seen without the counterexamples on the right.

Right: the right hand side is enough to communicate the answer; the right pattern can be seen without counterexamples on the left.

Flipping a BP will switch its sorting.

The following is taken from the comments on page BP513 (keyword left-narrow):

Call a pattern "narrow" if it is likely to be noticed in a collection of examples, without any counterexamples provided.

A collection of triangles will be recognized as such; "triangles" is a narrow pattern. A collection of non-triangular shapes will just be seen as "shapes"; "not triangles" is not narrow.

Narrow patterns tend to be phrased positively ("is [property]"), while non-narrow patterns opposite narrow patterns tend to be phrased negatively ("is not [property]").

See keywords left-narrow and right-narrow.

Adjacent-numbered pages:
BP825 BP826 BP827 BP828 BP829  *  BP831 BP832 BP833 BP834 BP835

dual, handed, leftright, meta (see left/right), miniproblems, contributepairs, viceversa, presentationinvariant

boxes_bpimage_three_per_side [smaller | same | bigger]

Aaron David Fairbanks