Electron-positron annihilation into hadrons is one of the fundamental tools for studying non-perturbative QCD effects. Fragmentation functions, describing the formation of hadrons from partons, are an indispensable tool in the interpretation of hadron-production data, e.g., in the investigation of nucleon structure via semi-inclusive deep-inelastic scattering. The cleanest process to access fragmentation functions is hadron production in electron-positron annihilation. However, little information can be derived on charge-separated fragmentation functions from single-inclusive hadron production. A better handle on the flavor contributions can be gotten by flavor correlations or tagging: the hadron type in one hemisphere puts constraints on the parton flavor in the other hemisphere and thus on the flavor decomposition of the hadronization process. This can be exploited in inclusive hadron-pair production in electron-positron annihilation. While two hadrons in the same hemispheres, e.g., originating from the same parton, open an avenue to an unusual class class of fragmentation functions, dihadron fragmentation functions, two hadrons in opposite hemispheres can be used for flavor and polarization tagging of single-hadron fragmentation functions. These scenarios have recently been subject to renewed studies at the Belle experiment. As one part of this contribution, without explicit relation to polarization effects, the dependences of the production cross section of pairs of identified light mesons (charged pions and kaons) as well as of (anti)protons on the individual z of the hadrons as well as on alternative scaling variables will be presented. Furthermore, the first measurement of the transverse-momentum dependence of single-hadron fragmentation, for charged pions, kaons as well as for protons, will be discussed.
The Belle experiment at the asymmetric KEKB electron-positron collider has already published several ground-breaking results on fragmentation functions, among others, the first measurement of azimuthal correlations for pairs of almost back-to-back charged pions related to the novel Collins fragmentation effect. The latter is one example of polarization dependence in hadronization despite an unpolarized initial state. Recently, these Collins modulations for charged pions have been revisited in the study of their transverse-momentum dependence. The analysis was also expanded to include for the first time neutral pions as well as eta mesons. An additional novel polarization effect is the formation of transversely polarized Lambda hyperons in the fragmentation process. Transversely polarized hyperons have been produced copiously in unpolarized proton-proton collisions, however, the question of the origin of these partially rather large polarizations has remained unanswered. One possibility is a novel fragmentation function, which was the focus of a recent study at Belle. The results for the production of transversely polarized Lambda hyperons and the of above-mentioned Collins modulations will be discussed in the second part of this contribution.